You are on page 1of 40

The International Journal of Flexible Manufacturing Systems.

2 (1990): 289-328
9 1990 Kluwer Academic Publishers, Boston. Manufactured in The Netherlands.

Flexibility in Manufacturing: A Survey

Bureau d'Economie Thdorique et Appliqu~e, Universitl Louis Pasteur, Strasbourg, France


Faculty of Management, Unversity of Toronto, Toronto, Canada

Abstract. This article is an attemptto surveythe vastliterature on flexibilityin manufacturingthat has accumulated
over the last 10 to 20 years. The survey begins with a brief review of the classical literature on fexibility in
economics and organizationtheory, which provides a background for manufacturing flexibility. Several kinds
of flexibilities in manufacturingare then defined carefully along with their purposes, the means to obtain them,
and some suggested measurements and valuations. Then we examine the interrelationships among the several
flexibilities. Variousempirical studies and analytical/optimizationmodelsdealing with these flexibilities are reported
and discussed. The article concludes with suggestions for some possible future research directions.

Key Words: Flexibility in manufacturing, discrete parts manufacturing, flexibility

1. Introduction

With the emergence of new microprocessor technologies, the concept of flexibility in

manufacturing has become a key consideration in the design, operation, and management
of manufacturing systems. A substantial amount of literature dealing with manufacturing
flexibility has accumulated over the last ten years. The major part of this literature is devoted
to defining various types of flexibilities and identifying systems that exhibit one or more
of these. Some papers also deal with the issues of the measurement and/or valuation of
the various flexibilities. According to Ettlie (1988), few rigorous systematic treatments of
the topic of flexibility in manufacturing, let alone empirical studies of actual manufacturing plants, have been reported that give a coherent statement of the strategic as well as
tactical implications of this important dimension of manufacturing strategy. The literature
makes one thing abundantly clear: flexibility is a complex, multidimensional, and hard-tocapture concept. At least 50 different terms for various types of flexibilities can be found
in the manufacturing literature. Usually, there are several terms referring to the same flexibility type. Definitions for these terms that have appeared in the literature are not always
precise and are, at times even for identical terms, not in agreement with one another (see
also Swamidass 1988). Not much work has been done to develop analytical models that
deal with the concepts of flexibility rigorously, and of course, to determine the optimal
levels of flexibility (see also Slack 1987). As a result, the measures proposed in the literature
are naive and, at times, somewhat arbitrary.
Moreover, the management of flexibility remains poorly understood. According to
Jaikumar (1987), "With few exceptions, the flexible manufacturing systems installed in the




United States show an astonishing lack of flexibility [when compared to their Japanese
counterparts]. In many cases, they perform worse than the conventional technology they
replace. The technology itself is not to blame; it is the management that makes the difference" (see also Adler 1985; the Economic Commission for Europe 1986). Baranson
(1983) argues that the global view held by Japanese firms towards marketing and production explains why the managers there take a long-term and comprehensive view towards
capital investments that considers not only cost savings in labor, material, and space, but
more significantly the broader strategic implications of increased flexibility (to respond
to changes in consumer demands and competitive threats) and versatility (in meeting diversifted market demands) in designing and producing of products (see also De Meyer et al.
1989). Empirical evidence also supports the view that flexibility does not get its proper
due at the time of decision making with regard to investment in manufacturing technology
(Lim 1987; Krasa and Llerena 1987).
The purpose of this article is to document the evolution of our understanding of the concept of flexibility in manufacturing. In order to do this, we shall first review briefly the
economic and organizational literature on flexibility that dates back to the early 1920s and
the late 1950s, respectively. We shall then survey the substantial literature on manufacturing flexibility, most of which has accumulated over the last ten years since the advent of
the flexible manufacturing system (FMS). We shall be concerned mainly with discrete parts
manufacturing, which includes job shops, assembly lines, flexible transfer lines, and FMSs
including flexible assembly systems. To accomplish this task, we shall try to classify various
flexibilities that have appeared in the literature and organize our survey around them. Our
purpose here is not to develop a detailed taxonomy. Rather, it is to facilitate an overview
of the various flexibilities and their interrelationships that have been reported in the literature.
The plan of the article is as follows. In the next section, we provide a historical perspective on the topic. Section 3 describes the concept of manufacturing flexibility and its stra~gic
importance in general terms. Definitions of specific flexibilities along with suggested
measurements and interrelationships between them are developed in section 4. Empirical
studies attempting to compare flexibilities of different manufacturing systems and those
making international comparisons in this context are briefly discussed in section 5. The
article concludes with some remarks and suggested future research in section 6 and an
extensive bibliography containing relevant references.

2. Flexibility: A historic perspective

Concern about flexibility is certainly not new. It has arisen in numerous economic and
organizational contexts in the last 70 years.

2.1. The economic view

An excellent early discussion appears in Lavington (1921), who draws a connection between random changes and the value of flexibility by considering the "risk arising from
the immobility of invested resources." later, in the context of the theory of the firm, Stigler




(1939) considers a plant to be flexible if it has a relatively flat average cost curve. Marschak
and Nelson (1962) argue that Stigler's notion of flexibility varies inversely with the slope
of the marginal cost curve; see the discussion of volume flexibility in section 4.7.3. This
also means that minimum average costs vary inversely with flexibility, or as Stigler put
it, "flexibility will not be a 'free good': A plant certain to operate x units of output per
week will surely have lower costs at that output than will a plant designed to be passably
efficient from x/2 to 2x units per week." By a simple example, Marschak and Nelson conclude that the relative desirability of a flexible plant (i.e., volume flexibility) increases as
the variation in market price (as measured by variance) increases and as the ability to predict
market price before making an output decision increases. Mills (1984) takes these ideas
one step further and shows how endogenous flexibility is determined in competitive markets
with demand fluctuations.
Hart (1940) recognizes that the postponement of decisions until more information comes
in, that is to say, the preservation of flexibility, is a fundamental means of meeting future
uncertainty (see also Tintner 1941). That individuals might have a preference for postponement of choice in the absence of risk and uncertainty is explored by Koopmans (1964).
Klein and Meckling (1958) regard the process of research and development of a new product as one in which the developer gradually acquires knowledge about the difficulty of
alternative ways of completing this task and makes a sequence of decisions, each allocating
a new part of his budget and each appropriate to the knowledge so far accumulated. Mass~
(1968) formulates a problem of choosing between rigid and flexible capital investments
by taking into account explicitly the cost of adoption of these investments to changes in
future environment.
Citing the works of Stigler, Hart, Klein and Meckling, and others for the notion that
good current actions may be those that permit good later responses to later observations,
Marschak and Nelson (1962) attempt to broaden the conjecture that flexibility is good in
an uncertain world. They propose three ordinal measures of flexibility in two-stage decision problems. One of these is the following: A first-stage action al is more flexible than
another action a2 if the set of possible second-stage decisions following al includes that
following a2. Furthermore, they show that the more the decision maker expects to learn
about the future stages of the world as time goes by, the more it pays him to concern himself
with the flexibility of his early actions. They also claim the complement--the greater the
flexibility in decision making, the greater the value of information-gathering. This converse issue is treated in detail in Merkhofer (1977). Kreps (1979) demonstrates that preferences
for flexibility can be treated axiomatically and shows that they may be equivalent to the
preferences derived from expected utility theory. Formalization of the notion of flexibility
in a sequential decision context and relating its value to the amount of information that
the decision maker expects to receive have been attempted by Rosenhead et al. (1972), Henry
(1974), Mandelbaum (1978), Jones and Ostroy (1984), Miller (1986), and others.
Rosenhead et al. defined and measured flexibility by the number of optional alternatives
left over after one has made an initial decision. Henry showed that if a model is simplified
by replacing all random variables with their means, then the simplified model may more
readily choose an inflexible "irreversible decision" than the original model might. Miller
shows that the relaxation of the assumption of independent demand in standard inventory
models leads to ordering of less inventories (i.e., more flexibility).



Mandelbaum defines flexibility as "the ability to respond effectively to changing circumstances" and observes that it can be characterized into two different forms: action flexibility, "the capacity for taking new action to meet new circumstance," and state flexibility,
"the capacity to continue functioning effectively despite changes in the environment."
Kulatilaka and Marks (1988) use a game theory formulation to value the (action) flexibility of being able to shift between labor-intensive and capital-intensive technologies. They
point to strategic advantages or disadvantages arising in bargaining with suppliers of labor
in the presence of incomplete contracts. They assume a world of certainty in order to isolate
the strategic aspects of flexibility from the option aspects. Vives (1989) shows in an oligopoly
context with incomplete information that the technological choice of the firm has both a
flexibility value and a strategic commitment value. Moreover, he studies the impact of increases in uncertainty on these values.
Jones and Ostroy (1984) consider explicitly the cost of switching from one action in this
period to another in the next. Their analysis brings out an important behavioral principle:
The more variable are a decision maker's beliefs, the more flexible is the position he will
choose. They emphasize: "The way flexibility is used to exploit forthcoming information
may be dictated by attitudes toward risk; but flexible positions are attractive not because
they are safe stores of value, but because they are good stores of options"
Not surprisingly, therefore, following Black and Scholes (1973), there have been attempts
to compute the value of flexibility, viewed as a hedge against future uncertainty, using their
option price formula; see, e.g., Andreou (1988), Triantis and Hodder (1989), He and Pindyck
(1989), and Richard (1989).
Because our discussion of flexibility as an economic, decision-theoretic concept has been
brief, we refer the reader for further details to Maier (1982) for a survey of flexibility articles in the European economic literature and to Cohendet and Llerena (1989) for a recent
comprehensive survey. To conclude our discussion, we leave the reader with the following
remarks from Jones and Ostroy (1984) to ponder.
In their paper, Jones and Ostroy indicate that there has been a long tradition of isolated
recognition that flexibility choice is a component of a wide range of economic decisions.
They surmise that its limited role in conventional microeconomic theory is perhaps due
to the difficulties of defining flexibility in a way that has universal application and of obtaining formal results without model-specific qualifications.

2.2. The organizational view

There is a substantial literature dealing with the concept of flexibility in an organizational
context. Feibleman and Friend (1945) define organizationalflexibility as the ability of an
organization to suffer limited change without severe disorganization; see also Reich (1932).
Ashby (1956) has proposed the law of requisite variety, which stipulates that the organization be complex in proportion to the complexity of the external stimuli it must deal with.
March and Simon (1958) have introduced the concept of organizational slack that provides
an organization with the excess resources to cope with internal as well as some environmental
uncertainties. Burns and Stalker's (1961) organic structure (as opposed to mechanistic structure), Emery and Frist's (1962) sociotechnical system, Walton's (1980) high commitment



systems, and some forms of decentralized, divisionalized, project management, and matrix
structures (see, e.g., Child 1982) refer to models of organization that have the flexibility
to operate responsively in a rapidly changing environment; see also Pugh et al. (1968),
Daft (1978), and Mintzberg's (1979) concept of ad-hocracy.
Recently, especially in the context of flexible technologies, new organizational forms
have been evolving beyond the traditional hierarchical or functional structure. One class
of organizational arrangements that is capable of much faster response to changing environment than functional structures may be called product-focused forms. These are organized
around the output functions rather than around the input functions that characterize traditional functional organizations. Each is organized "with the grain" to accomplish the particular tasks at hand (Lindholm 1975). These arrangements complement many of the flexible technologies. They have newer names, such as group technology cells, parallel assembly
cells, product verkstad (a Swedish term that literally means product shops and could be
translated as flexible focused factories), plants-within-plants, and network organizations;
see Kolodny (1989) for further details. Preece (1986) has defined a concept of structural
flexibility, which is concerned with the extent to which the structure of an organization
facilitates or hinders responsiveness of members of the organization to change. This change
could be initiated from within the organization itself or it could be a reactive change in
response to changes in the economic, social, or political environment of the organization.
Another specific concept of labor flexibility has been developed by the Institute of Manpower Studies (U.K.) (see, e.g., Atkinson 1985). Three main types of labor flexibilities
have been identified. Numericalflexibility concerns the readiness with which the number
of people employed can be adjusted to meet fluctuation in the level of demand; functional
flexibility concerns the readiness with which the tasks performed by workers can be changed
in response to varying business demands; financialflexibility is the extent to which compensation practices encourage and support the other two flexibilities that the firm seeks.
Following Feibleman and Friend (1945), Kozan (1982) defines work-groupflexibility as the
group's ability to adjust its activities to changing conditions without these adjustments
resulting in disorganization. He also develops and tests a measure for it in steady-state
The work on flexibility at the level of individual behavior can be seen in Rokeach (1960)
and Harvey et al. (1961). The latter authors view cognitively complex people as more flexible. People who are flexible are more able to deal with conflict and ambiguity.

2.3. The manufacturing context

Finally, we come to the history of flexibility in the context of manufacturing. Diebold (1952)
recognized flexibility to be essential for medium and short-run manufacturing of discrete
parts. As a break from the traditional philosophy of machine design that had the product
rather than the operation in view, Leaver and Brown (1946) and Diebold suggested machine
designs in terms of functions to be performed. Leaver and Brown (1946), in a fascinating,
ahead-of-its-time article, proposed a series of small, functionally oriented machines that
could be "plugged" together. Considering their design to be economically unjustifiable,
Diebold proposed his own concept of a machine that can simultaneously perform a bundle



offunctions that are related. Furthermore, it is noteworthy to mention that Diebold also
envisioned a concept that is reminiscent of flexible manufacturing as we know it today.
He wrote, "If we could couple a group of production machines, or similar machines designed
around the bundle of functions concept, by some form of inexpensive and flexible material
handling equipment, and add a control mechanism to do the work normally done by the
operator, we would have a factory completely automatic in terms of direct operation, although
there would still be need for considerable indirect labor"
Needless to say, these designs remained largely on the drawing board until the advent
of microprocessor technology. In practical terms, flexibility was viewed as a tradeoff against
efficiency in production and dependability in the marketplace (see, e.g., Abernathy 1978;
Wheelwright 1981; Hayes and Wheelwright 1984). The extreme situations of job shop being flexible but inefficient and Detroit-type mass production (Groover 1987) or automated
transfer lines being efficient but inflexible are well known in the literature. How to extend
flexibility to large-scale production without sacrificing efficiency was not known until the
late 1960s. Earlier in the 1950s, Herbert Simon (see Simon 1977, p. 24), recognizing that
humans are more flexible than machines (see also Diebold 1952), raised two questions:
1. What are the prospects for matching human flexibility with automatic devices?
2. What are the prospects for matching human skills in particular activities by reducing
the need for flexibility?
Simon goes on to say, "The second question is a familiar one throughout the history of
mechanization; the first alternative is more novel."
Simon relates the second question to the principle of homeostatic control of the environment, i.e., environmental control as a substitute for flexibility. He cites some examples
of this principle in work. One is the smooth road, which provides a constant environment
for the vehicle, thus eliminating the advantages of flexible legs. Another, more relevant
example for our purpose is that of automated transfer lines. In these lines, work in process
is presented, by means of transfer mechanisms, to successive machine tools in proper position
to be grasped and worked, eliminating the sensory and manipulative functions of workers
who formerly loaded such tools by hand. Thus, according to Simon, "We see that mechanization has more often proceeded by eliminating the need for human flexibility--replacing
rough terrain with a smooth environment--than by imitating it."
In contrast with mechanization, the development of FMSs beginning in the early 1970s
provided Simon's novel first alternative as well as Diebold's bold vision. This implementation is part of a larger movement of duplicating the capabilities of the sensory organs, the
manipulative organs, the locomotive organs, and the central nervous system. Simon has
conjectured that "automation of the functions wholly within the central nervous system
will be feasible long before automation of comparably flexible sensory, manipulative or
locomotive functions" In the arena of flexible manufacturing, we see that the automation
of symbol manufacturing, and perhaps thinking, has already happened. The automation
of the more complex eye-brain-hand sequences seems to have just begun; see section 4.9
on program flexibility defined by Jaikumar (1984).
With flexible manufacturing, it becomes possible to bring the efficiency of mass production to batch production of multiple products. Instead of economies of scale, the efficiency



in batch production is captured by the term economies of scope (Panzar and Willig 1981;
Goldhar and Jelinek 1983; Talaysum et al. 1986; see also Baumol et al. 1977). The efficiency of the midvolume, midvariety production is largely accomplished by a drastic reduction or elimination of setup costs and times required for switching from the production
of one product to another.
In the next section, we describe the concept of manufacturing flexibility in more detail.

3. The concept of manufacturing flexibility

Flexibility of a system is its adaptability to a wide range of possible environments that it
may encounter. A flexible system must be capable of changing in order to deal with a changing environment. According to Kickert (1985), flexibility can be considered as a form of
metacontrol aimed at increasing control capacity by means of an increase in variety, speed,
and amount of responses as a reaction to uncertain future environmental developments.
Flexibility in manufacturing means being able to reconfigure manufacturing resources
so as to produce efficiently different products of acceptable quality. An earlier definition
goes back to Ropohl (1967); he considers manufacturing flexibility as the property of the
system elements that are integrally designed and linked to each other in order to allow
the adaptation of production equipments to various production tasks.
Jaikumar (1984) emphasizes the fact that flexibility in manufacturing is always constrained
within a domain (see also Jubin 1981; Besson 1983; Goldhar and Jelinek 1983; and Gerwin
1989). Such a domain should be defined in terms of portfolio of products, process, and
procedures and should be well understood by product designers, manufacturing engineers,
and software programmers. The domain should be planned, managed, and with learning
expanded (see also Amendola and Gaffard 1986).
Moreover, there are other limitations on manufacturing flexibility that must also be defined. These include the speed and the cost of response (Riebel 1954; Swoboda 1964;
Gustavsson 1985; Garrett 1986), the amount of required reinvestment (Tarondeau 1986),
and the extent of interruptions in the existing system (Fiore 1984).
With regard to environmental uncertainties, it should be understood that manufacturing
flexibility is required in order for a firm to cope with both internal changes and external
forces (Garrett 1986). The internal disturbances for which flexibility is useful include equipment breakdowns, variable task times, queueing delays, rejects, and rework (Buzacott and
Mandelbaum 1985). External forces refer largely to the fundamental uncertainties of the
competitive environment (Behrbohm 1985; Zelenovic 1982; Garrett 1986; Maier 1982).
These uncertainties may be current or potential. Moreover, their probabilistic nature may
not always be known. Uncertainty may exist for level of demand, product prices, product
mix, and availability of resources. Uncertainty may arise out of actions of competitors,
changing consumer preferences, technological innovations, new regulations, etc.
Thus, manufacturing flexibility clearly has major implications for a firm's competitive
strength. This significant role of manufacturing flexibility makes it a part of the firm's
strategy. By strategy here, we mean "a set of plans and policies by which a company tries
to gain advantage over its competitiors" (Skinner 1985). Indeed, Hayes and Wheelwright
(1984) consider flexibility as one of the dimensions of the competitive strategy of a business



along with price (and therefore, cost), quality, and dependability. Furthermore, priorities
assigned to each of these dimensions determine how the business positions itself relative
to its competitors.
It follows therefore that decisions regarding manufacturing flexibility arise from strategic
considerations (see also Lim 1987). Its development requires considerable managerial attention and can no longer, as in the past, be relegated as technical detail. Hayes and
Wheelwright (1984) go even further when they say that the potential of manufacturing in
general as a competitive weapon and the concept of using manufacturing as a strategic
asset have been almost always overlooked by management. According to Skinner (1985),
it is not always easy to grasp the interrelationship between manufacturing operations and
corporate strategy. What is required is the concept of manufacturing strategy, which in
the words of Hayes and Wheelwright "consists of a sequence of decisions that, over time,
enables a business to achieve a desired manufacturing structure (i.e, capacity, facilities,
technology, and vertical integration), infrastructure (i.e., workforce, quality, production
planning/material control, and organization), and a set of specific capabilities (that enables
it to pursue its chosen competitive strategy over the long term)."
Manufacturing flexibility must, therefore, be a permanent preoccupation (Behrbohm 1985;
Maier 1982) and not just an improvisation (see also Eversheim and Schaefer 1980; Becker
1985). It is much more than simply buying an FMS (Garrett 1986). The idea that flexibility cannot just be bought but must be planned and managed is a crucial one (Beste 1958;
Meffert 1969; Rempp 1982; Jaikumar 1984; Gustavsson 1985; Ranta 1988; Stecke 1989).
Management of manufacturing flexibility must invariably come to terms with the question of what are the "optimal" levels of various types of flexibilities. The answer to this
question requires that the management identify and be able to measure the various flexibilities that the manufacturing system must have in order to gain maximum competitive
advantage (see also, e.g., Gerwin 1987; Ettlie 1988; Ranta and Alabian 1988). This is certainly a difficult question, and we address it in the next section.

4. Def'mitions, purposes, means, and measurements of various flexibilities

This section is the heart of our survey. Here, we carefully define several different kinds
of flexibilities that are reported in the literature. Our definitions may not be identical to
those existing in the literature. Modification is required because terminology is not standardized, and in some cases, the definitions of particular flexibilities that exist in the literature
do not agree. Having settled on the definitions of the various flexibilities, we will be able
to discuss each of them in terms of its purposes, operational as well as strategic, the means
to obtain it, and suggested measurements and/or valuation. While the purposes of a flexibility express why it is needed, the means refer to the firm's technological and managerial
responses to that need.
In all, we shall organize our discussion under 11 different flexibilities in the first 11 subsections. These are machine, material handling, operation, process, product, routing,
volume, expansion, program, production, and market flexibilities. As we shall see, the
first three of these refer to flexibilities of the important components of the system, i.e.,
machines, material handling system, and the parts to be produced, respectively. The




















Figure 1. Linkages between the various flexibilities,

remaining flexibilities apply to the manufacturing system as a whole. Figure 1 provides

a convenient overview of the various flexibilities under consideration; see section 4.12 for
a further explanation of the figure. The reference that we shall follow most closely is that
of Browne, Dubois, Rathmill, Sethi, and Stecke (1984), although we deviate from their
view occasionally.
In section 4.12, we summarize briefly the linkages between the various flexibilities. Ranking of typical systems with respect to some of the flexibilities is provided in section 4.13.
Finally, in section 4.14, we review briefly the models that attempt to characterize optimal
flexibility choices.
Before we begin defining these flexibilities, we should note that a sophisticated computer
and information technology and a flexible organizational structure underlie each of them,
both at the component and at the system levels (see figure 1). It is because of this technology
that flexibility in manufacturing has become possible without a considerable sacrifice in
efficiency. Generally speaking, a cellular architecture with distributed information seems
to be most favored (see, e.g., McLean et al. 1983; Ranky 1986; Shaw and Whinston 1988).
While Kusiak (1986) has defined the concept of computer system flexibility measured by
its adaptability to the changing functions, we shall not do so. Rather, we choose to indicate
the computer hardware/software requirements during the discussion of individual flexibilities.



Also, these flexibilities cannot attain their full potential without the support of an appropriate
organizational structure. While the notions of labor and organizational flexibilities exist
in the literature (see section 2), a detailed treatment of this vast subject is neither possible
nor appropriate in this survey. Our preference here is to indicate particular organizational
considerations whenever needed in the discussion of individual flexibilities. Moreover, the
organizational issues are also brought out in our review of selected empirical studies in
section 5.

4.1. Machine flexibility

Machine flexibility (of a machine) refers to the various types of operations that the machine
can perform without requiring a prohibitive effort in switching from one operation to another.
In assembly systems, the term machine refers usually to an assembly robot. Types of
operations that we have in mind are drilling holes up to ~A"diameter, grinding case-hardened
steel to specified tolerances, assembling parts of certain shapes and sizes, etc. Note that
we allow these operations to include the specification of the input material, such as its
hardness or ductility. With regards to the prohibitive effort, it is usually expressed in terms
of time and cost. It would certainly rule out redesigning the machine completely. On the
other hand, it might not exclude changing the tools in the tool magazine.
This definition, while not contradictory, is different from the one provided by Browne
et al. (1984). It is closer to that of Carter (1986), who defines it to be the universe of possible uses of the machine and the ease of converting from one use to another; see also Riebel
(1954) and equipment flexibility of Son and Park (1987). Scharf's Einsatzflexibilitaet (1975)
emphasizes the former part of the definition. Bergner's Rueso'lexibilitaet (1979) emphasizes
the latter part (see also Maier 1982; Tarondeau 1982). Tarondeau (1986) defines separately
the term input flexibility to emphasize the variations in the input stock the machine can
The motivation for our definition lies in how useful the definition is in assessing the
contribution of the given machine toward the manufacturing flexibility of a system, not
yet fully specified and subject to major changes in the long run, of which the machine
will become an element. Such an assessment may enable us, for example, to make an informed decision to buy or not to buy the machine (see also Gerwin 1987).

4.LL Purposes. According to Ranta (1989), "The machine level provides the basic framework
for flexibility. Software functions cannot help to provide any extra flexibility, if the machines
are hard and expensive to change" In other words, machine flexibility is necessary for
other flexibilities. At its own level, machine flexibility allows lower batch sizes (Ranta 1988)
and resulting savings in inventory costs, higher machine utilizations (Hutchinson 1984),
production of complex parts (Ranta and Alabian 1988), shorter lead times for new product
introductions (Carter 1986), and better product quality realizations in the face of random
variations in input quality (Tarondeau 1986).

4.1.2. Means. Technological sources of machine flexibility are numerical control, easily
accessible programs, rule-based languages, sophisticated part-loading and tool-changing



devices to ensure easy changeability of workpieces and tools, size of the tool magazine,
availability of sufficient pallets and fixtures, number of axes, automatic chip removal,
adaptive control to optimize metal removal, diagnostic software, integration with CAD/CAM,
etc. Thus multipurpose, multiaxis adaptable CNC machining centers are highly machine
flexible. Lim (1987) has studied 12 firms with different FMS designs and concluded that
the weakest part in machine flexibility seemed to be the unavailability of automated fixture
assembly and mounting; see also Ranta (1988). According to Jaikumar (1984), group
technologies attempt to improve machine flexibility, in contrast to FMSs, which improve
process flexibility as defined in section 4.4.
Machine flexibility requires considerable attention on the part of the management.
Operators need to be trained to acquire programming, maintenance, and diagnostic skills.
Quality circle activities along with the authority of workers to stop production as in Japan
(Gerwin 1989) can result in gradual changes that increase machine flexibility. For this,
machines must be installed in a way so as to avoid physical limitations that would inhibit
these changes (Ranta 1988).
Regarding what the future holds for machine flexibility, "the big problem," according
to Ranta (1989), "still is to integrate basic tooling functions, turning, drilling and milling,
into a universal machining center . . . . Electronic and software development will yield completely new prospects in this respect--it might help to create real (machine) flexibility in
a cost-efficient manner . . . . One radical innovation which would change the whole picture
is laser processing. If the technical reliability of lasers increases, they could become an
effective means of increasing flexibility (milling, drilling and turning by the same tool;
no tool maintenance and drift; flexibility of software; applicability to different materials)."
In the context of assembly systems, Boothroyd (1982) describes a Universal Assembly
Center. Here, since a universal (generic or flexible) gripper that will grip any part would
be prohibitively expensive (see also Ranta 1989), one requirement for such a center is that
the parts have been designed so that they can be fed in one of the "programmable" feeders
and gripped by the "universal" gripper.

4.L3. Measurements. The first aspect of machine flexibility can be measured by the number
of different operations that a machine can perform without requiring more than a specified
amount of effort. Brill and Mandelbaum (1987) suggest a measure weighted over a given
set of tasks. The weights reflect the relative importance of the tasks and the effectiveness
with which the machine can perform them (see also Carter 1986). To measure the second
aspects, several authors have emphasized the effort in terms of time and/or cost required in
switching from one operation to another (see, e.g., Riebel 1954; Swoboda 1964; Wildemann
1977). Son and Park (1987) measure it in terms of the opportunity of the machine to add
value to raw materials or, more specifically, by the ratio of the total output and the idle
cost of the machine for a given period.
Other measures include the number of tools or the number of programs (Tarondeau 1982)
that the machine can use, the extent of variations in key dimensional and metallurgical
properties of the raw input stock the machine can handle (Gerwin 1987), and the rate at
which the machine becomes obsolete when a new product is introduced. This last measure,
suggested by Gustavsson (1984) and Lam (1987), can be expressed as



Investment's residual value for the desired new model
Original investment in the machine

Of course, it must be noted that this measure depends, in general, on what the new product is.

4.2. Material handling flexibility

Flexibility of a material handling system is its ability to move different part types efficiently
for proper positioning and processing through the manufacturing facility it serves.
The definition covers loading and unloading of parts, transporting them from machine to
machine, and eventually storing them under varying conditions of the manufacturing facility.
This definition is consistent with the discussion in Diebold (1952), Stecke and Browne
(1985), Eidenmueller (1986), Peter (1984), Kusiak (1986), and Chatterjee et al. (1984, 1987).
The latter authors define the ability of the material handling system in terms of physical
location of each group of machines, the linkages between each pair of groups and between
each pair of machines within each group, and the times for every possible move between
machines. From these it is possible to define the set of all possible material paths that can
be supported in the factory. Peter emphasizes buffer sizes, the ability to accommodate different parts of different shapes and sizes, and the readjustment of paths in case of expansion. The definition also subsumes the pallet fixture flexibility defined by Newman (1986).
This flexibility determines the degree of freedom available to part loading schedules.
4.2.L Purposes. Material handling flexibility is very important for various system flexibilities under consideration. Having a flexible material handling system increases availability
of machines and thus their utilization and reduces throughput times. According to Rattner
et al. (1988), material handling robots and automated storage and retrieval systems increase
the information processing capabilities of the production system.
4.2.2. Means. Material handling flexibility can be attained by having transporting devices
such as forklift trucks and push carts and an appropriate layout design. In highly automated
facilities, devices such as automated guided vehicles, robots, and computer control, which
can send parts to new paths in cases of blocking and machine breakdowns, would be needed
to acquire material handling flexibility. Having a number of general-purpose fixtures will
also increase the flexibility. Newman (1986) describes a system that allows approximately
350 different parts to be mounted on only four different fixture attachments, which in turn
are mounted on a general-purpose fixture. Handling the part once it arrives at the machine
is also important. Thus, automatic tool changers, multiaxis robots, etc. will enhance material
handling flexibility.
As we had mentioned in section 4.1.2, robots with flexible grippers and intelligent interfaces (tactile sensors, vision, signal processing) are at present prohibitively expensive. According to Ranta (1989), real flexible grippers and reliable tactile sensors may be commercially available only after 1995 at reasonable prices.
Drawing on impressions from work organization designs in Sweden, Kolodny (1985) emphasizes better layouts, more space, cleaner environment, better ergonomics, and the use of
autonomous work groups to improve material handling flexibility in assembly environments.



Parallel-assembly arrangements allow an assembler not to be impeded by a person who

might be slower or might be experiencing a problem. Changes such as local testing and
inspection in the assembly process increase the cycle time and may considerably expand
the required skills of assemblers as they must learn to test, inspect, adjust, and repair products, as well as to assemble them. This may call for work-group arrangements that group
several operators and their material into a physically bounded area. Various manual and
motor-propelled jigs and fixtures rotate parts and assemblies through several axes and adjust
to the needs of a range of assemblers. This means improvements in ergonomics and a better
health and safety environment. With personal access of materials, operators in a work group
become more independent of staff and systems and are more able to plan and control their
own work activities. This provides a better opportunity for flexibility than in the traditional
dyad consisting of a superior and a subordinate.

4.2.3. Measurements. Chatterjee et. al. (1987) define a universal material handling system
that can link every machine to every other machine. Then the material handling flexibility
of a given system can be expressed by the ratio of the number of paths that the system
can support to the number of paths supported by the universal system. Note that this ratio
also gives an indication of the inhibition of manufacturing flexibilities, especially the routing
flexibility to be defined later, due to the material handling system (see also Das and
Khumawala 1989).
Stecke and Browne (1985) have ranked the following systems in order of increasing flexibility: belt conveyors, powered roller conveyors, power-and-free conveyors, monotractors
or monorails, towlines, and automated guided vehicle systems. While their emphasis is
on the evaluation of various material handling devices in terms of their influence on different
flexibilities proposed in Browne et al. (1984) and on the overall production flexibility,
Klarhorst (1981) and Knipschild (1986) identify important characteristics of the devices and
weight these characteristics to obtain a measure of their material handling flexibility. Newman
(1986) has estimated the increase in FMS performance that would result by increasing
material handling flexibility through the use of general-purpose fixtures.

4.3. Operation flexibility

Operation flexibility of a part refers to its ability to be produced in different ways.
Operation flexibility is a property of the part, and means that the part can be produced
with alternate process plans, where a process plan means a sequence of operations required
to produce the part. An alternative process plan may be obtained by either an interchange
or a substitution of certain operations by others. Thus, a part that permits operations to
be performed in alternate orders or using different operations (i.e., slurry versus wirebrush deburr) in an interchangeable fashion would possess operation flexibility.
A process will be considered to have operation flexibility if parts that are being produced in the system possess operation flexbility and if the material handling system is able
to deliver parts to machines in different possible orders. The definition of the operation
flexibility of a process is consistent with Browne et al. (1984), Maier (1982), and Chatterjee
et al. (1987).



4.3.1. Purposes. Operation flexibility of parts contributes to various system flexibilities,

especially the routing flexibility. Operation flexibility of a process allows for easier scheduling of parts in real time (Browne et al. 1984) and increases machine availability and utilization, especially when machines are unreliable.
4.3.2. Means. Operation flexibility of a part derives from its design; see Maier (1982) for
an example. The design should allow the parts to have surfaces that are easily accessible
for various operations. Parts that are assembled from standardized components or parts
that are modular (Gustavsson 1985; Besson 1983; Ranta 1988) are likely to exhibit operation flexibility. Systems such as CAD/CAM, computer-aided process planning (CAPP),
and group technology make it easier to design parts possessing operation flexibility.
4.3.3. Measurements. Operation flexibility of a part can be measured by the number of
different processing plans for its fabrication.

4.4. Process flexibility

Process flexibility of a manufacturing system relates to the set of part types that the system
can produce without major setups.
This definition is similar to the one in Browne et al. (1984). Another preferred term is
mix flexibility used by Gerwin (1982) and Carter (1986). Buzacott (1982) uses the term job
flexibility, which relates to the ability of the system to cope with changes in jobs to be
processed by the system; Rempp (1982) and Behrbohm (1985) use the term Einsatzflexibilitaet; Freist et al. (1984), Kegg (1984), and Melcher and Booth (1987) use the termpartmix flexibility; and Yamashina et al. (1986) call it variant flexibility. Falkner (1986) considers a system to be process-flexible if the manufacturing costs are relatively stable over
widely ranging product mixes. The use of the term short-term flexibility by Warnecke and
Steinhilper (1982) emphasizes the set of parts that can be produced in the short run.
4.4.L Purposes. The main purpose of process flexibility is to reduce batch sizes and reduce
inventory costs (Browne et al. 1984; Ranta and Alabian 1988). This can be accomplished
even when there are shifts in the product mix demanded by the market; Carter (1986) refers
to this purpose as that of insurance in the short term (see also Slack 1987). Carter also
emphasizes that process flexibility allows machines to be shared and thus minimizes the
need for duplicate or redundant machines. Process flexibility, according to Gerwin (1989),
satisfies the strategic need of being simultaneously able to offer to customers a range of
product lines.
4.4.2. Means. Process flexibility of a system derives from the machine flexibility of
machines, operation flexibility of parts, and the flexibility of the material handling system
composing the system. Ranta (1988) emphasizes the need for supporting planning flexibility
along with machine flexibility in ensuring process flexibility. Multiskilled workers who
can handle different products and the ability to transfer a variety of fixtures and tooling
into and out of the system enhance process flexibility (Gerwin 1989).



4.4.3. Measurements and valuation. An obvious measurement would be the volume of

the set of part types that the system can produce without major setups. One could perhaps
use group technology concepts to define the set of part types. Volume may be expressed
by the number of different part types in the set (Browne et al. 1984; Ancelin 1986; Gerwin
1987) if they can be counted, and if not, by the range of sizes, shapes, etc. (Proth 1982);
see also Lasserre and Roubellat (1985) for a volume measure in another context, namely
decision flexibility. All tooling, fixtures, and other manufacturing resources must be available
within the manufacturing system and must allow for possible extremes of mix variations.
Jaikumar (1986) and Ettlie (1988), in their surveys of FMSs, asked the firms to count the
number of part types produced in their FMSs as a measure of process flexibility. Carter
(1986) proposes to measure it by the extent to which product mix can be changed while
maintaining efficient production. Warnecke and Steinhilper (1982) measure it by the
changeover cost between known production tasks within the current production program.
Ettlie (1988) uses average changeover time as a second measure of process flexibility in
his FMS survey. Carter suggests that the average changeover time should be viewed in comparison to average cycle time of machines. Son and Park 0987) measures it by the ratio
of the total output and the waiting cost of parts processed for a given period.
Buzacott (1982) focuses on the infeasibility components, i.e., parts that the process cannot produce. He uses the set of all jobs that cannot be processed by the system multiplied
by the probability that such a job in fact will be required to be processed in the future.
However, Jaikumar (1984) has pointed out that without restricting the domain of interest,
the set of parts that cannot be processed would be arbitrarily large compared with the set
of parts that can be processed. Thus, the notion of infeasibility is difficult to measure.
Moreover, any FMS is designed for a restricted domain of a particular family of parts.
Jaikumar argues that the family of parts intended for the FMS to produce should be viable
in the long run. While the choice of such a family is very important, one needs only to
consider flexibility within such a domain once it is chosen.
Jaikumar (1984) measures process flexibility by the expected values of a defined portfolio of products that can be processed through the system of limited resources for a given
set of contingencies. He formulates a stochastic mathematical program from which one
can derive "shadow costs" of the reduction of contingencies, such as machines and tools
being unavailable (see also Fine and Freund 1988; Gupta, Buzacott, and Gerchak 1988).
The motivation behind his formulation is an attempt to measure the market value of process flexibility. Moreover, Jaikumar insists that this valuation is an underestimate, since it
does not take into account the value of "the skill generated within the people working with
the system and the advantages of the management philosophy and culture which goes with
it;' which can often be quite large. Mandelbaum and Buzacott (1986) use the decision theory
approach described briefly in section 2 to derive the value of having a process-flexible system
as opposed to an inflexible system. Recently, Andreou 0988) has suggested the use of the
option price formula of Black and Scholes (1972) to compute the value of process flexibility in a given situation. Triantis and Hodder (1989) incorporate downward-sloping marginal
profit curves for products in the firm's product mix and a production capacity constraint.
They also allow downward-sloping demand curves for the underlying assets constructed
as proxies for the manufacturing cost structure. This results in complex exercise decisions
for the options or the contingent claims which comprise the value of the production system.



In the case of two products and two underlying uncertainties, Triantis and Hodder derive
an explicit formula for the value of the process flexible production system; see also He
and Pindyck (1989).

4.5. Product flexibility

Product flexibility is the ease with which new parts can be added or substituted for existing
In other words, product flexibility is the ease with which the part mix currently being
produced can be changed inexpensively and rapidly. It should be kept in mind that the
addition of new parts will invariably involve some setup. This distinguishes product flexibility from process flexibility. What is required for product flexibility is that the setup
does not involve inordinate amounts of time and cost. We should also emphasize that the
new parts in the definition above cannot be arbitrary; for further clarification, see section
4.10 on production flexibility. It is important to note that in Lim's (1987) survey of FMSs
in the United Kingdom, 11 out of 12 reporting companies considered flexibility in manufacturing to mean product flexibility.
The definition is consistent with Browne et al. (1984), Gerwin's (1982) and Falkner's
(1986) parts flexibility, Mandelbaum's (1978) action flexibility, Zelenovic's (1982) design
adequacy, the dynamic flexibility of Cohendet and Llerena (1987), short product lifeflexibility of Yamashina et al. (1986), and parts-change flexibility of Freist et al. (1984) and
Kegg (1984). This definition also subsumes Gerwin's (1982) design change flexibility, as
well as the changeover and modification flexibilities of Gerwin and Tarondeau (1989).
Hedrich (1983) and Maier (1982) define a related concept of Fertigungsredundanz, by which
they mean a system with a built-in redundancy to carry out tasks that are currently not needed.

4.5.L Purposes. Product flexibility allows the company to be responsive to the market by
enabling it to bring newly designed products quickly to the market (Carter 1986; Gerwin
and Tarondeau 1989). Since the future product designs are usually unknown, it becomes
important to design and develop the production facility to be product-flexible. According
to Hayes and Schmenner (1978), smaller companies in many industries often adopt a strategy
of competing on the basis of product flexibility, i.e., their ability to handle difficult, nonstandard orders and to lead in new product introduction. It should be noted that Tombak (1988)
in an extensive econometric study finds that product flexibility is more important in the
growth phase of a product than in its mature phase. Therefore, in the markets that are rapidly
in flux due to short and uncertain product life cycles, product flexibility along with a
sophisticated computer-aided design capability provides the company with a formidable
competitive weapon.

4.5.2. Means. Product flexibility depends on machine flexibility, material handling flexibility, operation flexibility, efficient CAD/CAM interface, CAPP, group technology organization, use of similar part programming routines, rapid exchange of tool and dies, flexible
fixtures, etc. Gerwin (1989) suggests keeping the amount of hard tooling to a minimum
and offlining conversion of a part of the system if required, so that the rest of the system



continues to operate. Moreover, software should be designed so that it can readily be changed
when new products appear in order to facilitate rerouting of products while some of the
work stations are being converted off-line (see also Dogan and Davis 1989). Ranta (1988)
advocates modular system software for accomplishing such tasks. Gustavsson (1984) links
product flexibility to the manufacture of products assembled from standardized parts. Here
the product is differentiated only in the later stages of its production. This can be compared with Tarondeau's (1982) diffdrenciation retard~e; see also Benassy et al. (1986) and
Hollard and Margirier (1986).
Jaikumar (1984) emphasizes the incorporation of systematic learning obtained from the
production of product in the current portfolio to nurture the product flexibility of the system.
That means workers must be willing and able to continually learn new operating procedures
(Gerwin 1989).

4.5.3. Measurements and valuation. Product flexibility can be measured by time or cost
required to switch from one part mix to another, not necessarily of the same part types
(Browne et al. 1984; Buzacott 1982; Zelenovic 1982; Warnecke and Steinhilper 1982).
Hollard and Margirier (1986) suggest that the above cost should be expressed in relation
to the total production cost. Indeed, Son and Park (1987) measure it by the ratio of total
output to setup costs for a given period. In his comparison of FMSs in Japan and the U.S.,
Jaikumar (1986) uses the number of new parts introduced per year as one of the measures
of product flexibility; see also Gerwin (1987). Jaikumar (1984) emphasizes the benefit aspects
of product flexibility by measuring it in terms of total incremental value of new products
that can be fabricated within the system for a defined cost of new fixtures, tools, and part
programs. This value can be obtained by certain shadow prices in an appropriately formulated stochastic mathematical programming problem. Kulatilaka (1988) also develops a
stochastic program that solves for the value of product flexibility together with the dynamic
operating schedule of the production process. Furthermore, he proposes to modify existing
capital budgeting techniques to incorporate special features of flexibility.
Triantis and Hodder (1989) indicate how their model of obtaining the option value of
process flexibility (see section 4.4.3) can be extended to include switching or setup costs
associated with adjusting the product mix. This extension would then provide us with the
option value for product flexibility.

4.6 Routing flexibility

Routing flexibility of a manufacturing system is its ability to produce a part by alternate
routes through the system.
Alternate routes may use different machines, different operations, or different sequences
of operations. Typically, these different machines (e.g., lathe and milling machines or two
brands of grinders) are those capable of essentially the same processes. It should be noted
that routing flexibility is different from operation flexibility in the sense that the former
is the property of a system while the latter is that of a part. Even a part with a single specified
operations sequence, i.e., no operation flexibility, may still be processed using different
routes through the system. It is also different from the material handling flexibility, which



is the property of a specific component of the system. Thus, with the existing material
handling subsystem, only some of the routes, by which it is possible to produce a part
under a universal subsystem (see section 4.2.3), may be feasible.
The definition is similar to Falkner (1986), Freist et al. (1984), Kegg (1984), Buzacott's
(1982) scheduling flexibility, Jaikumar's (1984) process flexibility, and Durchlauffreizuegigkeit of Herrman and Pferdmenges (1980), Maier (1982), and Behrbohm (1985).
It is also consistent with Gerwin (1982), Buzacott (1982), and Browne et al. (1984), although
these authors emphasize the system's ability to reroute parts in case of a machine breakdown.

4.~I. Purposes. Routing flexibility allows for efficient scheduling of parts by better balancing of machine loads. Furthermore, it allows the system to continue producing a given
set of part types, perhaps at a reduced rate, when unanticipated events such as machine
breakdowns, late receipt of tools, a preemptive order of parts, or the discovery of a defective part occur (see also Gerwin and Tarondeau 1989). Thus, it contributes toward the
strategic need of meeting customer delivery times. Routing flexibility also facilitates capacity
expansion if needed (Ranta and Alabian 1988).
4.(~2. Means. Routing flexibility comes about by having multipurpose machines, machines
with overlapping process envelopes, pooling of identical machines into machine groups
(Stecke and Kim 1989), system control software, versatility of material handling system,
and operation flexibility of parts (see, e.g., Browne et al. 1984; Yao 1985). According to
Falkner (1986), some planned underutilization of machines (or, redundancy in machines)
is needed in order for the system to be able to be rescheduled and maintain the overall
production rate in case of a machine breakdown. Gray et al. (1988) and Zhou and Wysk
(1989) point out the importance of an effective, integrated tool management system, while
Gerwin (1989) indicates the need for software aids for rearranging production schedules
when necessary.
Schonberger (1982) emphasizes the importance of labor flexibility in his study of Japanese
firms. More specifically, work group structure facilitates cooperation needed in case of
machine breakdowns. Workers need to have an intimate knowledge of the system to prevent damage and to reroute production (see also Gerwin 1989).
A survey of 12 FMSs by Lim (1987) revealed that the systems had very little or no routing
flexibility. Lim (1987) writes:
The lack of routing flexibility reflected both the state of technology and management
requirements for the systems. Routing flexibility, whether potential or actual, would require both software and hardware capabilities which might be beyond the resources of
suppliers or the inclination of individual companies. Having different routes for each
part, for instance, would not only entail extra memory capacity but also require some
real-time reasoning power on the part of the supervisory computer (intelligence) in order
to "understand" the nature of a breakdown, "sort" out the next best alternative, "transmit"
the diagnostics and necessary instructions on the altered route and updated schedule
to appropriate sub-systems including the operative(s) and support personnel. Apart from
duplicating processing assignments, this flexibility would also require tooling and
machine tool redundancy, all of which meant substantially greater capital costs and longer
project time.



4.~3. Measurements. Several alternative measures for routing flexibility have been proposed in the literature. Obvious measures are the average number of possible ways in which
a part type can be processed in the given system (Chatterjee et al. 1987; Chung and Chen
1989) and the ratio of existing number to possible number of links between machines in
the given system (Carter 1986; Primrose and Leonard 1984). Another network-based
measure in terms of entropy has been suggested by Yao (1985), Yao and Pei (1987), and
Kumar (1986). Routing entropy is defined to be the entropy measure of the information
contained in the list of operations and machines from which the next operation and the
machine must be chosen. Note that this is a dynamic measure that changes over time. Chung
and Chen also suggest a measure for the strategic value of routing flexibility by percentage
reduction in total job completion time due to its presence when compared with use of fixed
Browne et al. (1984) and Ancelin (1986), have developed measures that emphasize the
system's ability to handle unanticipated events. These measure percentage decrease in
throughput because of a machine breakdown (Buzacott 1982; Browne et al. 1984) or the
cost of the production lost as a result of expediting a pre-emptive order.

4. 7. Volume flexibility
Volume flexibility of a manufacturing system is its ability to be operated profitably at different overall output levels.
Note that only feasible output levels are under consideration here; see section 4.8. This
definition is similar to the ones in Browne et al. (1984), Gerwin (1982), Maier (1982),
Behrbohm (1985), Freist et al. (1984), and Kegg (1984). It is also similar to the demand
flexibility of Son and Park (1987). Volume flexibility has some degree of interchangeability
with Slack's (1987) deliveryflexibility--the ability to change planned or assumed delivery

4.7.1. Purposes. Uncertainty in the level of demand impedes the strategic objective of increasing and maintaining market share. A case in point is the costly efforts of General
Motors during the mid-1980s to stimulate the market and maintain capacity in the face
of declining sales (Gerwin 1989). Volume flexibility permits the factory to adjust production upwards or downwards within wide limits. Hayes and Schmenner (1978) point out that
successful companies in cyclic industries like furniture often exhibit this trait. According
to Slack (1987), volume flexibility has two aspects: speed of response and range of variations, the former being useful in the short term and the latter in the long term.

4.7.2. Means. If we are to model costs by only fixed and (constant) variable cost components, a system with given fixed and variable costs is more volume-flexible than another
system that chooses to have relatively higher fixed cost in order to have relatively lower
variable cost. As we show in section 4.7.3., average manufacturing cost of the former system
will be less sensitive to volume changes than that of the latter. However, costs are usually
nonlinear, and there are important adjustment costs that are associated with volume changes.



In these cases, a highly automated FMS may be volume-flexible because it allows the firm
to produce without a large amount of labor, which is difficult and expensive to adjust both
downward and upward (see also Ollus and Mieskonen 1989); Gerwin 1989). Bylinsky (1983)
cites the case of the Yamazaki plant near Nagoya, Japan, employing only 215 people (in
contrast with 2500 in a conventional factory) and having a maximum capacity of turning
out $230 million worth of machine tools a year. In this plant, it is claimed that the sales
can be reduced to $80 million a year, if need be, without laying off workers. To quote
Bylinsky (1983), "the plant illustrates (yet) another aspect of economy of scope: with flexible automation, a manufacturer can economically shrink production capacity to match
lower market demand"
Gerwin (1989) suggests that workers must possess skills that can be used elsewhere when
production volume decreases. He also recommends excess modular capacity that remains
unused except after breakdown occurs. Then the high capacity facilitates a quick return
to normal production and in-process inventory levels. Ranta (1988) emphasizes the importance of subcontracting network, and Monden (1982) suggests a JIT approach for volume
4.7.3. Measurements a n d valuation. Browne et al. (1984) measure volume flexibility by

how small the volume can be for all part types together with the system still being run
profitably. This really measures only the downside volume flexibility. An obvious generalization would be to measure volume flexibility by the range of volumes in which the firm
can run profitably. Gerwin (1987) measures it by the ratio of average volume fluctuations
over a given period of time to the production capacity limit. Falkner (1986) suggests, as
a measure of volume flexibility, the stability of manufacturing costs over widely varying
levels of total production volume.
To further elaborate, let us consider the simple paradigm of fixed cost F and constant variable cost c per unit of production. Then the total cost TC(V) and the average cost AC(V) are
TC(IO = F + cV, AC(IO = (F + cV)/V,
where V denotes the volume of production. It is easy to see that the (negative) elasticity
of the average cost with respect to volume is F/(F + cV). The elasticity increases with
F and decreases with c.
Following Stigler (1939) and Marschak and Nelson (1962), if the total cost function TC(IO
is increasing and convex, then we can define volume flexibility by 1/TCw. Note here that
AC(I/) is a U-shaped curve. Also, in the special case of the quadratic total cost TC(V) =
F + cV + 1~/2~, the volume flexibility 1/TC~ = ~.
Ancelin (1986) proposes the amount of slack capacity as a measure of volume flexibility.
More specifically, his measure is the following Potential Requirement Ratio (PRR):


total available time - (required time + maintenance time)

required time

_ nonrequired time - maintenance time

required time



Thus, PRR represents a portion of the nonrequired time that can be mobilized for processing. Son and Park (1987) measure it by the ratio of the total output and the inventory/shortage
costs of finished products and raw materials for a given period.
Market valuation of volume flexibility can be obtained as the shadow price associated
with the constraint that the demand should be met in an appropriately formulated
mathematical programming problem. This measure is related to the one above in the sense
that a slack capacity implies a zero shadow price.

4.& Expansion flexibility

Expansion flexibility of a manufacturing system is the ease with which its capacity and
capability can be increased when needed.
The capacity is in terms of output rate per unit time, whereas capability refers to such
characteristics as quality, the technological state, other types of flexibilities, etc. Note that
in contrast with volume flexibility, expansion flexibility is concerned with capacity, i.e.,
the maximum feasible output level. Expansion flexibility makes it easier to replace or add
machinery by providing for such possibilities in the original design. Ease in this connection
refers to the overall effort needed for the expansion. It would include the direct cost, the
indirect cost of interruption in production because of the expansion, and the speed with
which the expansion can be accomplished.
This definition is similar to the one in Browne et al. (1984), Carter (1986), and Falkner
(1986). Buzacott and Mandelbaum (1985) also include the system's ability to contract as
a part of the definition. Related concepts are Zelenovic's (1982) design flexibility, Young
and Murray's (1986) system flexibility, and Jacob's (1974) Entwicklungsflexibilitaet.

4.&l. Purposes. Expansion flexibility is important for firms with growth strategies such
as venturing into new markets, since it permits step-by-step adaptation of the system for
expansion. In contrast, volume flexibility serves survival strategies such as maintaining
existing markets and profitability. According to Carter (1986), expansion flexibility helps
to reduce implementation time and cost for new products, variations of existing products,
or added capacity.

4.&2. Means. Expansion flexibility can be achieved in several ways, such as by building
small production units (Buzacott and Mandelbaum 1985), having modular flexible manufacturing cells (Browne et al. 1984; Burstein 1986), having multipurpose machinery that does
not require special foundation and a material handling system that can be more easily routed
(Carter 1986; Lim 1986), having a high level of automation that can facilitate mounting
additional shifts, providing infrastructure to support growth, and planning for change (or,
planned flexibility; see Eversheim and Schaefer 1978).
A critical technical issue for the future expansions is the possibility of a module-type
control structure and a transportation device that allows for a soft extendability of the system
without drastic architectural changes. Hall and Stecke (1986) suggest the use of automated
guided vehicles as transportation devices to support expansion flexibility, especially in flexible assembly systems. The system architecture (interfaces, etc.) should be one that enables
systems to be extended in a stepwise manner and new features to be added without major
new design efforts (Ranta 1988).



4.&3. Measurements. Expansion flexibility can be measured by the overall effort and cost
needed in terms of time to add a given amount of capacity (Carter 1986). More specifically,
one can measure it by the ratio of the cost of doubling the output of the system to its original
cost. Browne et al. (1984), on the other hand, suggest the measure of how large the system
can become. While this measure alone is not entirely satisfactory, some idea of the upper
bound on the amount of expansion is certainly useful.
Jacob (1974) has suggested a measure for expansion flexibility that indicates the degree
to which the expected long-run profit from the given system approaches an "ideal" profit.
More specifically, let G and N denote, respectively, the expected long-run profit of the
given system and of a non-expansion-flexible system, where the expectation is carried out
over the possible states of the system during the specified long run. Let I denote the expected long-run profit associated with the hypothetical scenario in which, for each state,
a corresponding optimal manufacturing system is in use. With these definitions, we can
define Jacob's measure for expansion flexibility F e as


G 9

Note that N < G < I so that 0 < Fe < oo. A slightly easier and perhaps more preferred
measure might be the ratio


0 < - - < 1 .
Note that the above-proposed ratio is also linear in G.

4.9. Program flexibility

Program flexibility is the ability of the system to run virtually untended for a long enough
This flexibility is an example of the state flexibility of Mandelbaum (1978) and of implementation of Simon's (1977) first alternative; see section 2. Also, it subsumes Zelenovic's
(1982) notion of adaptive flexibility. The definition is a slight modification of program flexibility defined by Jaikumar (1984). He uses the words "during the second and third shifts"
instead of "for a long enough period" in the above definition. Untended operation during
the second and third shifts has been found to be extremely important by Jaikumar (1984,
1986) in his comparison study of American and Japanese FMSs. In such a mode of operation, inspection, fixturing, and maintenance can be performed during the first shift.

4.9.L Purposes. According to Jaikumar (1984), program flexibility reduces the throughput
time by having reduced setup times, improved inspection and gauging, and better fixtures
and tools. Increased understanding of the process required to attain program flexibility results
in procedures that can produce products with tighter tolerances and better quality. Being
able to work untended increases the effective capacity of the production system. Thus program flexibility allows simultaneous improvement of productivity and quality.



4.9.2. Means. Program flexibility depends on process and routing flexibilities and on having sensors and computer controls for detection and handling of unanticipated problems
such as tool breakages, part flow jams, etc. Newman (1986) includes quality control and
tool maintenance (see also Gray et al. 1986; Zhou and Wysk 1989) as important elements.
Program flexibility is a stringent requirement, since it necessitates a thorough understanding of the system; only then can all of the contingencies of the system be proceduralized.
Jaikumar (1984) suggests a new management paradigm that involves managing knowledge
as a resource. This requires controlled experimentation for learning, modification of the
accumulated knowledge in the form of software programs, and the ability to transfer the
knowledge among similar products, processes, and procedures. Ollus and Mieskonen (1989)
emphasize the need for a flexible organization such as those described in section 2.

4.9.3. Measurements and valuation. An obvious measure would be the expected percentage uptime during the second and third shifts. Bright's (1958) scale of measuring automation level ranging from 1 (lowest) to 17 (highest) might also be usable here. A measure
of market valuation is suggested by Jaikumar (1984). He formulates a stochastic mathematical
programming problem with emphasis on reduction of throughput times and increase in
the effective system capacity in untended manufacturing and proposes to measure program
flexibility by certain shadow prices that arise in the mathematical programming formulation.

4.10. Production flexiblity

Production flexibility is the universe of part types that the manufacturing system can produce without adding major capital equipment.
Minor resources such as new tools may be allowed in order to define the relevant universe.
In contrast with product flexibility, note that production flexibility may allow considerable
setups but not major capital equipment. In a way, production flexibility defines the set of
existing and potential (or even hypothetical) parts, from which the candidate parts can be
drawn as the new parts considered in the definition of product flexibility. In Slack's (1987)
terminology, product flexibility is a response flexibility, whereas production flexibility is
a range flexibility; see also the long-term flexibility of Warnecke and Steinhilper (1982).
This definition is similar to those proposed by Browne et al. (1984) and Carter (1986).
The concept of production flexibility can also be related to the variable structure property
of Ropohl (1971), the state flexibility of Mandelbaum (1978) and Buzacott (1982), the productflexibility of Fine and Freund (1978), the fexible manufacturing module flexibility of
Kusiak (1986), and the factory system flexibility of Freist et al. (1984) and Kegg (1984).
Hayes and Wheelwright (1984) indicate that production flexibility might be thought of as
a third dimension to the product process matrix; see Stecke and Raman (1986) for a further
development of this idea.

4.10.L Purposes. Production flexibility allows the firm to compete in a market where new
products are frequently demanded. Production flexibility minimizes the implementation
time for new products or major modifications of existing products (Carter 1986). On the
operational level, it permits an increase of part families and allows the firm to diversify
its risk.



4.10.2. Means. As Browne et al, (1984), Carter (1986), and Chatterjee et al. (1987) imply,
production flexibility depends on the variety and the versatility of the machines that are
available, the flexibility of the material handling system in use, and the factory information
and control system. Thus, the production flexibility derives from the capability of aggregation
of the flexibilities of the machines and material handling systems. The system can perform
all the operations that are inherent in the machines of the system. But, in terms of part
types that the system can make, production flexibility is more than the sum of the parts
that each individual machine can make. Of course, the presence of this synergy assumes
that a flexible material handling system is in place.
Production flexibility is related to the properties of the transportation system, warehousing system, interfacing system, distributed data bases, systems control, and software
modularity. An open communication system as well as the use of a common communication protocol will help to increase production flexibility (Ranta 1988). Slack (1987) emphasizes the importance of labor flexibility in obtaining production flexibility.
4.10.3. Measurement. An obvious measure for production flexibility is the size of the universe
of parts the system is capable of producing (see, e.g., Chatterjee et al. 1984). However,
since production flexibility is, in some sense, a "long-run product flexibility" its valuation
can be obtained in a fashion similar to that used by Jaikumar (1984) to determine product
flexibility. That is, it can be measured by certain shadow prices in an appropriate
mathematical programming problem formulated for a longer term.

4.11. Market flexibility

Market flexibility is the ease with which the manufacturing system can adapt to a changing
market environment.
This concept emphasizes the importance of market orientation in manufacturing. Especially in rapidly changing markets, the interface between production and marketing functions becomes crucial. It should be obvious that market flexibility of the manufacturing
system complements its production and program flexibilities.
Gerwin and Tarondeau (1988) refer to product, process, and volume flexibilities as marketorientedflexibilities. In addition, they include modification flexibility to allow for uncertainties that exist at the time of product design as to which product attributes customers
desire. With modification flexibility, a potential exists for implementing minor design changes
in a given product. Market flexibility subsumes the distribution flexibility of Ranta (1988),
which includes the inventory, transport, and administrative means of creating flexibility
in place, time, size, and assortment of deliveries to achieve customer satisfaction.
4.11.1. Purposes. Market flexibility is important for a firm's survival in environments that
are constantly in flux. Environments change because of rapid technological innovations,
change in customer tastes, short product life cycles, uncertainty in sources of supply, etc.
(see also Hutchinson and Holland 1982; Fine and Li 1988). Market flexibility allows the
firm to respond to these changes without seriously jeopardizing the business. Additionally,
market flexibility enables the firm to cash in on new business opportunities before its less



flexible competitors are able to. Market flexibility is essential if the firm's market strategy
emphasizes customized products and frequent product changes (see, e.g., Goldhar and
Jelinek 1983).

4.11.2. Means. As markets change, the manufacturing system may be required to process
new products, cope with fluctuating production volumes, and even to undergo capacity
changes. Thus product, volume, and expansion flexibilities contribute to market flexibility. Market flexibility requires that the process of production planning and inventory controis be closely integrated with such marketing functions as market forecasts, product
development, and customer relations. Moreover, good relationship with suppliers and welldeveloped distribution channels are also essential for market flexibility. An example would
be a successful implementation of an MRP II class software with user intervention (see
also Dogan and Davis 1989).
4.11.3. Measurements. Market flexibility can be expressed as a weighted measure of efforts in terms of time and cost required to introduce a new product, to increase and decrease
production volume by a specified amount, and to add a unit of capacity. Market flexibility
can also be measured by the shortage cost or the cost of delay in meeting the customer
orders (see, e.g., Abadie et al. 1988).

4.12. Linkages between various flexibilities

Having described various flexibilities in detail, we note that figure 1 summarizes the linkages
that have been reported to exist between them; see also Ollus and Mieskonen (1989) for
a similar figure. The figure indicates that flexibilities of components contribute to the various
flexibilities of the system. These in turn influence the aggregate flexibilities as shown. Viewed
from another perspective, the firm's manufacturing strategy dictates the extent of system
flexibilities and, in turn, of component flexibilities that the firm must possess. The figure
also indicates that the structure of organization and microprocessor technology underlies
all of the flexibilities.
Figures indicating the need for flexibilities to deal with variety and uncertainty in the
short and long terms and a hierarchy between flexibilities are also given in Slack (1987);
see also Browne et al. (1984) and Yilmaz and Davis (1987). Slack emphasizes that the
response aspect of a flexibility is needed in the short run while the range aspect is needed
in the long run. Yilmaz and Davis characterize flexiblities in terms of three attributes; at
times, after a time, and over time. According to them, machine and routing flexibilities
can be related to flexibility at times; operating, process, and product flexibilities can be
related to flexibility after a time; and volume, expansion, and production flexibilities can
be related to flexibility over time.

4.13. Ranking of flexibilities for different types of automated manufacturing systems

In this section, we review some attempts to ratak different types of automated manufacturing systems in terms of some of their flexibilities. Gerwin (1987) argues for usefulness of
these attempts in evaluation and selection of manufacturing processes. Abdel-Malek and



Wolf (1989) propose a method of ranking alternative FMS designs that is based on weighting
relevant attributes of the components making up the respective designs. Furthermore, they
indicate a need for a proper data base on the measurement of these attributes for different
components. The Kearney and Trecker Corporation (1982) positions the various systems
that it offers on the well-known volume/variety chart developed by the corporation.
Herrmann and Pferdemenges (1980), Maier (1982), and Behrbohm (1985) have ranked
different types of single NC-machine systems and multiple linked NC-machine systems
in terms of various manufacturing flexibilities. They have considered only two types of
multiple machine systems, namely, FMSs with complementary machines and flexible transfer
lines. Other types of automation that should be considered are FMSs with interchangeable
machines and automated Detroit-type assembly lines (see also Falkner 1986). We have chosen
not to reproduce their assessments because of differences in their definitions of flexibilities
and rating scales used. There seems, however, to be a general agreement that in order of
increasing flexibilities (process, product, routing, and volume), the multiple-machine
automation systems can be ranked as follows: Detroit-type mass production, flexible transfer
line, complementary-type FMS, and interchangeable-type FMS. The ranking of the first
two and FMSs should be quite obvious. For some support of the ranking between the two
FMS types, see Falkner (1986). Stecke and Raman (1986) have attempted a more detailed
ranking; see also Kusiak (1986) and Miller (1985).

4.14. Optimization models concerning manufacturing flexibility

Before we conclude section 4, we review models that address the issue of manufacturing
flexibility in an optimization context. In view of a recent survey of these models by Fine
(1989), our discussion will be very brief. Models not dealt with by Fine are also included
in the following review.
Fine classifies the models under consideration in three groups based on the economic
phenomena they consider: 1) flexibility as a hedge against uncertainty; 2) interactions between flexibility and inventory; and 3) flexibility as a strategic variable that influences competitor's actions. We follow the same classification for our review.
The papers in group 1 purport to value process or product flexibilities and have already
been mentioned in sections 4.4.3 and 4.5.3. In group 2, Porteus (1985) extends the economic
order quantity framework to include investment to reduce setup costs. Chand and Sethi (1989)
examine the effect of learning in setup reduction in the dynamic lot size model. Karmarkar
and Kekre (1987) use an economic production quantity type model to provide conditions
under which owning a flexible machine is better than owning several dedicated machines.
The tradeoff is between more expensive dedicated capacity and the holding costs due to
cycle stocks on account of the changeover downtime (see also Vander Veen and Jordan
1988). Graves (1988) presents a model to analyze interaction between safety stocks and
process and volume flexibilities; this model could be used to optimize inventory and flexibility. Caulkins and Fine (1988) show that in the case of a seasonal demand, flexible capacity
and inventory can be either substitutes or complements depending on the pattern of the
demand. Milgrom and Roberts (1988) study substitution between inventories and communication with customers. They examine how the choice between these depends on such variables
as variability in demand, cost of expanding product lines, production and inventory costs,
and communication costs. A much more general model than the one in group 2 reviewed



by Fine is that of Milgrom and Roberts (1989a). They formalize the idea of complementary
groups of activities that exist between production, marketing, and engineering design and
incorporate it into a nondifferentiable, nonconvex profit-maximization model of a
monopolistic (or a monopolistically competitive) firm. The defining characteristic of these
groups of complements is that if the levels of any subsets of the activities are increased,
then the marginal return to increases in any or all of the remaining activities rises. A number
of flexibility aspects are represented in the model by way of including such variables as
number of product improvements per product per period, design cost per product improvement, setup cost on newly changed products, number of setups per period, and delivery
time. Milgrom and Roberts use purely algebraic (lattice-theoretic) methods to argue that
the profit optimization over time generates many of the observed patterns (or clustering
of characteristics) that mark modern manufacturing.
Marketing in these firms emphasizes low prices, high quality, and frequent product improvements. Production wants such things as low inventory levels and product designs that
use common inputs. At the organizational level, there is extensive use of close communication
and joint planning with suppliers. Moreover, nonconvexities in the model mitigate against
any smooth distribution of these characteristics among firms. Thus, any evidence of distinctly
separate clusters of firms characteristics--firms tend to broaden their product lines or tend
to become focused factories--would tend to support the Milgrom-Roberts theory.
In group 3, there are few game-theoretic papers that analyze the competitive dynamics
involving flexible manufacturing technology. Fine (1989) remarks that "considering the
wealth of observers who tout strategic benefits of flexible manufacturing systems, this shortage is a little surprising."
Gaimon (1988) considers a two-firm quadratic differential game model to compare how
the firms' technology acquisition strategies compare under the assumption of open-loop
and closed-loop dynamics. The results suggest that firms charge higher prices, acquire
less new technology, buy less total capacity, and earn higher profits in a closed-loop game.
Fine and Pappu (1988) assume two firms and two markets with each firm having a monopoly
in its own market. There are two possible equilibria. In one, both firms buy flexible technology and end up with lower duopoly profits. In the other, each firm buys flexible technology only for defensive purpose, i.e., each stays in its own markets. Once again, both firms
are worse off.
Milgrom and Roberts (1989b) extend their model described in group 2 above to incorporate game-theoretic aspects as well as uncertainties in the environment.
Finally, Roeller and Tombak (1989) explore the implications of market structure on the
selection of FMSs. They analyze an n-firm two-stage game in which firms choose between
a flexible and dedicated technology in the first stage, and then choose production quantities in the second stage. They find that as industries become more competitive, firms
diversify in technologies, i.e., FMS firms become less dominant while firms with dedicated
technologies become predominant.

5. Selected empirical studies

In this section, we review some of the empirical studies connected with flexibility. These
studies take the form either of case studies that deal with the measurement of flexibilities



for existing or hypothetical manufacturing systems or of surveys devoted to drawing some

general conclusions with respect to technological and management aspects of flexibility.
In section 5.1, we review the case studies while section 5.2 is devoted to reporting the results
from various surveys.

5.1. Case studies

Lim (1986) and I2tm (1988) have attempted to evaluate some existing manufacturing facilities
with respect to the flexibilities defined in Browne et al. (1984) and in Atkinson (1985).
Lim compares three companies that have implemented FMSs with different configurations
described in his paper. He specifies the reasons that led these companies to develop their
respective FMSs. These reasons are, respectively, new product line, response to market
requirement, and replacement of old machines. He also characterizes their respective products in terms of shape, size, weight, and material composition. He compares these facilities
with respect to the flexibilities defined in Browne et al. (1984). It should be noted, however,
that Lim's interpretation of the volume flexibility of Browne et al. (1984) as the ability
to produce different lot sizes is not correct. Lim also compares these facilities in terms
of their laborflexibilities defined by the Institute of Manpower Studies of the United Kingdom
and briefly described in section 2 of this article. He further concludes that all three systems
are very well developed in terms of machine flexibility but not in terms of production flexibility. Lam (1988) has assessed the degrees of various flexibilities of an FMS that produces printed circuit boards in a Northern Telecom plant. His assessment is carried out
on a scale ranging from low to medium to high.
Son and Park (1987) define two hypothetical manufacturing systems: a jobshop and an
FMS. They illustrate how they can obtain their quantitative economic measures for machine,
process, product, and volume flexibilities described in section 4. From these, they develop
an overall flexibility measure, which is the reciprocal of the sum of the reciprocal of the
individual flexibilities. It can also be expressed as the ratio of total output to the total cost
comprising the idle cost of machines, setup cost, waiting cost of parts being processed,
and inventory/shortage cost of finished products and raw materials for a given period. They
compute the overall flexibility of the FMS as well as the job shop and show that the FMS
has a higher overall flexibility even though the job shop has higher machine flexibility.
Moreover, they show that the FMS has higher productivity than the job shop. They emphasize that while there is an inverse relationship between productivity and each or some
of the individual flexibilities (Gustavsson 1984), there may still be a positive relationship
between productivity and overall flexibility (see also Chung and Chen 1989).
Finally, Graham (1986) considers two companies A and B that have purchased similar
turnkey FMSs from the same vendor at roughly the same time. The labor force of company A was rated to be technically skilled with a low workforce commitment. That of company B was rated to be lower in technical skills but higher in commitment. Graham noted
that both companies were having problems and that neither company could be rated to
be wholly flexible. Her conclusion was that companies A and B both failed to balance
and integrate the social and technical elements of the system. She suggested that future



research should focus on the questions of what in-house technical resources are needed
no matter how turnkey the project (see also Jaikumar 1984) and what organizational and
workforce choices fit better with the planning and operation of FMSs.

5.2. Surveys of FMSs and results

In this section, our task is to review empirical studies by Jaikumar (1986), Ettlie (1988),
Graham and Rosenthal (1986), Gerwin and Tarondeau (1989), Capdevielle and H~ran (1987),
Fix-Sterz et al. (1986), Sorge et al. (1983), Hollard et al. (1986), Lim (1987), Slack (1987),
Ollus and Mieskonen (1989), Tombak and De Meyer (1988), Tombak (1988), and Krasa
and Llerena (1987). We shall first describe some of the conclusions on which most of these
studies seem to agree. Subsequently, we shall point out important specific results suggested
in these studies.
International comparative studies show that the domestic culture may account for a significant portion of net national differences in the way FMSs are implemented and managed.
Management's understanding and attitudes toward the new technologies and the flexibilities
they permit play a crucial role. Yet the management seems to be lacking in understanding
of flexibility in the planning process of new technology projects (Lim 1987) as well as of
their implementation and of their management (Jaikumar 1986; Bocker et al. 1986).
Moreover, firms continue to use traditional evaluation tools, which do not account for flexibility, for making investments in flexible manufacturing (Krasa and Llerena 1987). This
is simply inadequate. A variety of suggestions and approaches to deal with the financial
justification of investments in flexible technology have been offered in the literature; see,
e.g., Myers (1984), Kaplan (1986), Krasa and Llerena (1987), I.ederer and Singhal (1988),
Hutchinson and Sinha (1989), Suresh and Meredith (1986), Monahan and Smunt (1987),
Buckiey (1987), Choobineh (1988), and Chand and Sethi (1982). The reader may also consult Wallace and Thuesen (1987) for an annotated bibliography on investment in flexible
The studies emphasize the importance of organizational and managerial changes to attain and enhance manufacturing flexibility. Flexible manufacturing makes it possible to
run batch production as a flow process. This changes both the content and the organization
of work. A new imperative is to manage process improvement, not just output. The challenge
is to develop physical and intellectual assets. This requires choosing projects that develop
these assets rather than monitoring the costs of day-to-day operations. Taylor's scientific
management must give way to new working methods. The organization chart of FMSs should
be flat, i.e., small, cohesive, self-regulating work groups consisting of highly skilled and
highly committed generalists should be organized around products, processes, and procedures (see also Maly 1988). This requires continuous training and learning in interpersonal and technical skills that would allow the transfer of knowledge obtained from working
with one product to another and continual improvement and enhancement of the FMS.
According to Jaikumar (1986), the use of small technologically proficient teams to design,
run, and improve FMSs represents a shift in focus from managing people to managing
knowledge and from production planning to project selection. Thus engineering acquires
the line responsibilities, while manufacturing becomes a service responsible for customizing its offerings to the preferences of special market segments.



Ettlie (1988) asked top and middle managers of 17 new manufacturing technology users
in the U.S. about their implementation strategies. He found out that process and product
flexibilities and the utilization rates of domestic systems were well above Jaikumar's (1986)
earlier results. Ettlie attributes this to domestic modernization that has recently brought
new systems with it that are much more flexible than reported earlier. He concludes that
generally firms adopting a manufacturing technology policy (MTP) simultaneously with
administrative innovations are more successful. By MTP, he means a firm's long-term plan
for modernization of the productive core. By administrative innovation, he refers to changes
in strategies, structures, and practices that may or may not be related to technological changes
in the operations or design core of an organization. More specifically, he found that flexibility is most likely to be significantly correlated with administrative experiments designed
to support customer integration. Ettlie points out further the importance of human resource
policy: it takes innovative people to be innovative, and firms have to recruit or develop
and keep these people fresh to support innovation. In this context, Ollus and Mieskonen
(1989) emphasize that the new skills necessary are not only concerned with technology,
but also with design, marketing, and management.
Gerwin and Tarondeau (1989) conducted a study of several French and American
automobile manufacturing companies. They focused on the work organization characteristics
of programmable automation and on the relationship of manufacturing flexibility to the
new technology. Because of the small amount of comparative data, the results are somewhat
tentative and discussion rather speculative. They show that American activities remained
mechanistic, reflecting the traditional belief that workers are a source of uncertainty and
not a resource. The French activities became more organic, reflecting the view that complex expensive technology requires operators to have higher skills, qualification, and training,
less task repetition, and more discretion (see also Capdevielle and H6ran 1987). It appears
that flexibility does not necessarily increase when programmable automation is introduced.
Whether gains or losses in flexibility occur depends upon the characteristics of the conventional process that is replaced. The average change in overall flexibility is negative at
low original automation levels and increases to become positive at high original automation levels. This result is consistent with the ISI-ISF report by Fix-Sterz et al. (1986).
Graham and Rosenthal (1986) emphasize that those organizations that exercise the
downgrading strategy of focusing on control and job polarization largely achieve shortterm flexibilities such as those of routing or process. More generally, their study of eight
FMSs in the U.S. shows that many companies did not understand the nature of their organizations in terms of installing integrated systems, which as long-term projects require time
for thought and study as well as highly committed, multiskilled people. Thus, to matrix-in
somebody, say from the systems resource group, to fulfill the need for a particular skill
in an integrated environment has typically meant handing that person an assignment without
a full-time, whole-hearted commitment (see also Maly 1988). Such persons report to several
different managers at the same time, and they detract from the effectiveness of the persons
who are there full-time. In short, continuity is crucial for the success of long-term
technological projects.
Hollard et al. (1986), in their survey on FMSs and flexible cells in France, found out
that automation of machinery and computerization of information do not increase the flexibility of the processes compared to job shop production but do reduce the time of information



processing and increase the computational capacity. In France as well as in West Germany
(see Fix-Sterz et al. 1986; Sorge et al. 1983), the programming and maintenance functions
are integrated at the shop-floor level. Sorge et al. (1983) also studied British firms and
found that owing to job specialization in Britain, programming in these firms was more
likely to be done by planners and production engineers than by operators. It should also
be noted that less job specialization exists in Germany due in part to the national training
Lim (1987) discusses his findings of a survey of 12 companies in Britain in which he
tried to identify strategic and operational objectives for implementing FMSs. He states
that on account of the lack of consistency (intracompany) and homogeneity (intercompany)
in the management's long-term and short-term objectives, the requirements for flexibility
were inconsistent or noncomplementary at best, and in conflict amongst themselves and
with the objectives at worst. The lack of clear understanding of flexibility alternatives at
both the strategic and operational levels might indicate that the design and introduction
of an FMS have not been carefully thought out (see also Krasa and Llerena 1987). Moreover,
it seems that the complexity of an FMS makes it easy to become submerged in technical
requirements, thereby overlooking the holistic view. This result is consistent with Slack
(1987), who made an empirical study in ten manufacturing companies on how managers
view the flexibility of their manufacturing system. He found out that managers focus more
on flexibility of individual resources (technology, labor, or infrastructure) than on the flexibility of the production system as a whole. Moreover, there is a tendency in most organizations to be more immediately concerned with response flexibility (short term) than with
range flexibility (long term). The failure to distinguish between range (of states a production system can adopt) and response (the ease with which it moves from one state to another,
in terms of cost, time, or organizational disruption) was a major cause of confusion between managers in the same organization.
Slack's study suggests that different types of manufacturing are concerned with the flexibility of different resources. For example, jobbing/batch manufacturers equate flexibility
with machine flexibility, while companies in the batch/mass production area seem to concentrate on labor flexibility. At the total manufacturing system level, managers identify
product, process, volume, and delivery flexibilities; see section 4.7. Slack also concluded
that managers seek to limit the need to be flexible (competing on a nonflexible basis, using
modular product design, and matching market segmentation with segmentation of the production system). Flexibility is seen as a means to other ends, rather than as an end in itself
(see also Ettlie 1988). It is judged in terms of how it enhances other measures of manufacturing performance, namely dependability, productivity, and availability. The types and
dimensions of the flexibility needed are determined by the variety of products, processes,
and activities, as well as by the ability of the system to cope with uncertainty.
Finally, using two data bases that cover 410 European and 168 U.S. firms, Tombak and
De Meyer (1988) find that managers use FMSs to accommodate variability in their inputs
and to enable them to produce a wider variety of outputs. Tombak (1988), based on a sample of 1455 business units from a PIMS database, developed a regression model that finds
flexibility te be an important explanatory variable of the performance of business units.



6. Conclusions
In this article, we have surveyed the relevant literature dealing with the concept of manufacturing flexibility. Because of the multidimensionality of this concept, several different types
of flexibilities have been defined in the literature. We have provided careful definitions of
some of these flexibilities, which we consider to be important, in order to clarify and to
survey the literature. We also hope that this work will contribute toward an eventual taxonomy or standardization of the flexibility terminology.
We have described various operational or raw measures and in some cases financial valuations of these flexibilities. Operational measures help manufacturing managers to understand the kind and extent of flexibility embedded in their production process and allow
them to make informal judgments on new equipment. Flexibility valuations can be incorporated in the capital budgeting analyses required for financial justification of the new
equipments under consideration.
We have also reviewed several analytical or optimization models dealing with decision
making with regard to various flexibility alternatives. Finally, we have reported on several
empirical studies assessing flexibility measures of real existing systems and examining the
related technological and organizational aspects.
Several future research directions can be discerned from our survey of the literature.
Clearly, there is a need for a detailed taxonomy of manufacturing flexibilities. With regard
to measurement, it may be possible to collect data on raw measures and to use such statistical
techniques as factor analysis to come up with a suitable flexibility scale (Gerwin 1987).
Factor analysis may identify the minimum number of significant hypothetical concepts that
underlie raw measures. It may then be possible to develop desiredflexibility profiles according to which vendors may be asked to design equipments under consideration.
It is also important to explore the tradeoffs that exist between various flexibilities and such
other factors as productivity, quality, and degree of automation. Increased control of one
dimension opens new opportunities of increased flexibility in other dimensions (Ettlie 1988).
With regard to optimization models, we should emphasize that flexibility should be viewed
as a competitive weapon in the arsenal of the firm's manufacturing strategy. In this role,
flexibility can be used to deal with uncertainties whose nature, especially their probabilistic
nature, may not even be known. Perhaps a game theory framework needs to be developed
within which the contribution of flexibility towards a firm's competitive strategy can be
assessed. Moreover, in the absence of knowledge regarding uncertainty, the idea of forecast
horizons, i.e., the time period beyond which forecast of future data may not have significant impact on current decisions, may need to be integrated in the game theory formulations as indicated in Sethi and Sorger (1989).
Finally, we recall from section 2 the remark by Jones and Ostroy (1984) about the limited
role of flexibility in economic theory because of its hard-to-capture nature. It is our hope
that manufacturing flexibility will not suffer the same fate.

An earlier version of this article was presented at the TIMS/ORSA Meeting, Washington,
DC, April 25-27, 1988.



This research is supported in part by NSERC under grant A4619 and by Manufacturing
Research Corporation of Ontario. Comments from Dave Quirin, Harvey Kolodny, Martin
Evans, Donald Gerwin, Kathy Stecke, and the referees are gratefully acknowledged. We
are also thankful to Scott Allerston and Tom Finlay of Faculty of Management Library
for helping us to obtain many of the cited references and to Connie Liew for typing several
versions of this paper. Finally, we are lovingly thankful to Chantal Angelina Sethi, our
other joint product, for her timely arrival, which permitted flexible scheduling in the writing
of this joint paper.

Abadie, J.P., Cohendet, E, H~mn, E, Krasa, A., and Llerena, E "R&luim les d~lais de r&action pour am~liorer
la production." Revue francaise de gestion, Vol. 67, pp. 91-102 (January-February 1988).
Abernathy, W.J. The Productivity Dilemma: Roadblock to Innovation in the Automobile Industry, The Johns Hopkins
University Press, Baltimore, MD (1978).
Abdel-Malek, L. and Wolf, C. "Evaluating Flexibility of Alternative FMS Designs--A Comparative Measure."
Working Paper, Division of Industrial and Management Engineering, New Jersey Institute of Technology, Newark,
NJ 07102 (1989).
Adler, P. "Managing Flexibility: A SelectiveReview of the Challenges of Managing the New Production Technologies
Potential for Flexibility." A Report to the Organization for Economic Cooperation and Development, Stanford
University, Palo Alto, CA 1985; also in California Management Review, Vol. 30, No. 3, pp. 34-56 (1988).
Amendola, M. and Gaffard, J.L. "Creation de technologie et recherche de flexibilitY: probl~mes et m~thodes
d'une approche analyfique." Working Paper, Universit~ de Nice, Nice, France (March 1986).
Ancelin, B. "Quels crit~res de performance pour les nouveaux ateliers?" Presented at the Journ~e d'~tudes Systimes de Production, Ecole des Mines de Paris, Paris, France (21 May 1986).
Andreou, S. "A Capital Budgeting Model for the Evaluation of Flexible Plant Capacity." Research Publication
GMR-6490, General Motors Research Lab., Warren, MI (1988).
Ashby, W.R. An Introduction to Cybernetics, John Wiley & Sons, Inc., New York, NY (1956).
Atkinson, J. "Flexibility: Planning for an Uncertain Future" Manpower Policy and Practice, Vol. 1, (Summer 1985).
Baranson, J. Robots in Manufacturing--Key to International Competitiveness, Lomond Publications Inc., Mt.
Airy, MD (1983).
Baumol, W.J., Bailey, E.E., and grdlig, R.D. "Weak Invisible Hand Theorems on the Sustainability of Multiproduct
Natural Monopoly." The American Economic Review, Vol. 67, No. 3, pp. 350-365 (1977).
Becker, H. "Einsatz der Kostenrechnung in mittelgrossen Industrieunternehmen--Eine empirische Untersuchung."
Zeitschriftfuer betriebswirtschaftliche Forschung, Vol. 37, pp. 601-617 (July-August 1985).
Behrbohm, P. Flexibilitaet in der industriellen Produktion, Peter Lang, Frankfurt/Main, West Germany (1985).
Beste, T. "Groessere Elastizitaet durch unternehmerische Planung vom Standpunkt der Wissenschaft" Zeitschrifi
fuer handelswissenschaftliche Forschung, Vol. 10, p. 75 (1958).
Benassy, J., Benchimol, G., Bloch, G., FerrY, A., Philip, C., Rostan, G., Sauvage, L., and Vayssier~, P. L'usine
intbgrbe par ordinateur, Hermes Publishing, Pads, France (1986).
Bergner, J. "Vorbereitung der Produktion, physische." in Handwoerterbuch der Produktionswirtschaft, W. Kern
(Ed.), Poeschel Verlag, Stuttgart (1979).
Besson, P. L'atelier de demain: perspectives de l'automatisationflexible, Presses universitaires de Lyon, Lyon,
France (1983).
Black, F. and Scholes, M. "The Pricing of Options and Corporate Liabilities,' Journal of Political Economy
Vol. 81, pp. 637-659 (1973).
Bocker, H.J., Daniels, J.P., Prasad, J.N., and Shane, H.M. "The Factory of Tomorrow: Challenges of the Future?'
Management International Review, Vol. 26, pp. 36-49 (1986).
Boothroyd, G. "Economics of Assembly Systems." Journal of Manufacturing Systems, Vol. 1, No. 1, pp. 111-127



Bright, J.C. '~utomation and Management" Division of Research, Graduate School of Business Administration,
Harvard University, Boston, MA (1958).
BriU, EH. and Mandelbaum, M. "On Measures of Flexibility in Manufacturing Systems?' W.P. # W87-08, University
of Windsor, Windsor, Ontario (June 1987).
Browne, J., Dubois, D., Rathmill, K., Sethi, S.P., and Stecke, K.E. "Classification of Flexible Manufacturing
Systems." The FMS Magazine (April 1984).
Buckley, J.J. "The Fuzzy Mathematics of Finance," Fuzzy Sets and Systems. Vol. 21, No. 3, pp. 257-273 (1987).
Burns, T. and Stalker, G.H. The Management of Innovation. Travistock Publications, London, UK (1961).
Burstein, M. "Finding the Economical Mix of Rigid and Flexible Automation for Manufacturing Systems?' In
Proceedings of the Second ORSA/TIMS Conference on Flexible Manufacturing Systems (Ann Arbor, MI), K.E.
Stecke and R. Suri (Eds.), Elsevier, Amsterdam, The Netherlands, pp. 95-106 (1986).
Buzacott, J. "The Fundamental Principles of Flexibility in Manufacturing Systems" Flexible Manufacturing Systems,
Proceedings of 1st International Conference (Brighton, UK), Elsevier, North Holland, Amsterdam (1982).
Buzacott, J. and Mandelbaum, M. "Flexibility and Productivity in Manufacturing Systems," in Proceedings of
the liE Conference, (Chicago, IL), Industrial Engineering and Management Press, Atlanta, CA, pp. 404-413 (1985).
Bylinsky, G. "The Race to the Automatic Factory." Fortune, pp. 52-61 (Feb. 21, 1983).
Capdevielle, P. and H~ran, E '~utomatisation, ~volution de l'organisation et des techniques et restructuration
des activit~s des ouvriers, des techniciens et des agents de maftrise?' Project Report, CEREQ, Strasbourg,
France (1987).
Carter, M.E "Designing Flexibility into Automated Manufacturing Systems" In Proceedings of the Second
ORSA/TIMS Conference on Flexible Manufacturing Systems (Ann Arbor, MI), K.E. Stecke and R. Suri (Eds.),
Elsevier, Amsterdam, The Netherlands, pp. 107-118 (1986).
Caulkins, J. and Fine, C. "Seasonal Inventories and the Use of Product-Flexible Manufacturing Technology."
Working Paper, O.R. Center, MIT, Cambridge, MA (1988).
Chand, S. and Sethi, S. ' ~ Dynamic Lot Size Model with Learning in Setups?' Forthcoming in Operations Research.
Chand, S. and Sethi, S. "Planning Horizon Procedures for Machine Replacement Models with Several Possible
Replacement Alternatives?' Naval Research Logistics Quarterly, Vol. 29, No. 3, pp. 483-493 (1982).
Chatterjee, A., Cohen, M.A., and Maxwell, W.L. ' ~ Planning Framework for Flexible Manufacturing Systems?'
WP #87-07-04, University of Pennsylvania, Philadelphia, PA (July 1987).
Chatterjee, A., Cohen, M.A., Maxwell, W.L., and Miller, L.W. "Manufacturing Flexibility: Models and
Measurements?' In Proceedings of the First ORSA/T1MS Special Interest Conference on FMS (Ann Arbor, MI),
K.E. Stecke and R. Suri (Eds.), Elsevier, Amsterdam, The Netherlands (1984).
Child, J. "Culture, Contingency and Capitalism in the Cross Nation Study of Organizations." In Research in
Organizational Behavior 03. Staw and L.L. Cummings, Eds.)
Child, J. Organization: A Guide to Problems and Practice. Second edition, Harper & Row, London, UK (1982).
Chung, C.H. and Chen, I.J. '~. Systematic Assessment of the Value of Flexibility from FMS." In Proceedings
of the Third ORSA/TIMS Special Interest Conference on FMS. (Cambridge, MA), K.E. Stecke and R. Suri
(Eds.), Elsevier, Amsterdam, The Netherlands, pp. 27-34 (1989).
Choobineh, F. "Uncertainty Treatment in Economic Evaluation of Manufacturing Systems." Presented at the
ORSA/TIMS, Denver, CO, October 23-26, 1988.
Cohendet, P. and Llerena, P. "Flexibilities, Complexity and Integrations in Production Processes." W.P. #8704,
Universit~ Louis Pasteur, Strasbourg, France, (1987).
Cohendet, P. and Llerena, P. "Flexibilit~s, risque et incertitude dans la th~orie de la firme: un survey." In Flexibilitb, information et dbcision, P. Cohendet and P. Llerena (Eds.), Economica, Paris, France, pp. 7-71 (1989).
Daft, R.L. ' ~ Dual-Core Model Organizational Innovation," Academy of Management Journal. Vol. 21, pp. 193-210
Das, S.R. and Khumawala, B.M. "Flexible Manufacturing Systems: A Production Management Perspective."
Production and Inventory Management Journal, Vol. 30, No. 2, pp. 63-67 (Second Quarter 1989).
Diebold, J., Automation: The Advent of the Automated Factory, Van Nostrand, New York, NY (1952).
De Meyer, A., Nakane, J., Miller, J., and Ferdows, K. "Flexibility: The Next Competitive Battle, The Manufacturing Futures Survey." Strategic Management Journal, Vol. 10, pp. 135-144 (1989).
Dogan, C.A. and Davis, R.P. ' ~ Software Design Model for Tactical Decision Making in a Manufacturing Environment." In Proceedings of the International Industrial Engineering Conference (Toronto, Canada), Institute
of Industrial Engineers, Atlanta, GA, pp. 610-615 (1989).



Economic Commission for Europe. Recent Trends in Flexible Manufacturing. United Nations, New York, NY (1986).
Eidenmueller, B. "Neue Planungs- und Steuerungskonzepte bei flexibler Serienfertigung" Zeitschriftfuer die
Betriebswirtschaftliche Forschung, Vol. 38, pp. 618-634 (1986).
Emery, E and Trist, E.L. "Socio-Technical Systems?' In Management Science Models and Techniques. C.W.
Churchman and M. Verhurst (Eds.), Pergamon Press, London, pp. 83-97 (1960).
Ettlie, J.E. "Implementation Strategies for Discrete Parts Manufacturing Technologies?' Final Report, The University
of Michigan, Ann Arbor, MI (April 1988).
Eversheim, W. and Schaefer, F.W. "Planung des Flexibilitaetsbedarfs yon Industrieunternehmen." Die Betriebswirtschaft. Vol. 40, pp. 229-248 (1978).
Falkner, C.H. "Flexibility in Manufacturing Plants." In Proceedings of the Second ORSA/TIMS Conference on
Flexible Manufacturing Systems (Ann Arbor, MI), K.E. Stecke and R. Suri (Eds.). Elsevier, Amsterdam, The
Netherlands, pp. 95-106 (1986).
Fine, C.H. "Development in Manufacturing Technology and Economic Evalutation Models." In Logistics of Production and Inventory. S. Graves, A. Rinnooy Kan, and E Zipkin (Eds.). North Holland, Amsterdam, The
Netherlands (1990).
Fine, C.H. and Freund, R.M. "Economic Analysis of Product-Flexible Manufacturing Systems Investment Decisions." In Proceedings of the Second ORSA/TIMS Conference on Flexible Manufacturing Systems (Ann Arbor,
MI). K.E. Stecke and R. Suri (Eds.). Elsevier, Amsterdam, The Netherlands, pp. 55-68 (1986).
Fine, C. and Li, L. "Technology Choice, Product Life Cycles, and Flexible Automation." Journal of Manufacturing and Operation Management, Vol. 1, No. 4 (1988).
Fine, C. and Pappu, S. "Flexible Manufacturing Technology and Product-Market Competition." Working Paper,
O.R. Center, MIT, Cambridge, MA (1988).
Fiore, C. "Le r61e de la logistique dans les syst~mes productiques: les impacts sur les investissements de PME?'
Presented at the Journke Productique, Montpellier, France (June 1984).
Fix-Sterz, J., Lay, G., Schultz-Wild, R., and Wengel, J. "Flexible Fertigungssysteme und -zellen in Rahmen neuer
Fabrikstrukturen." Project Report, Fraunhofer Institut fuer Systemtechnik und Innovationsforschung (ISI),
Karlsruhe, and Institut fuer Sozialwissenchaftliche Forschung (ISF), Munich, West Germany (1986).
Friest, C., Granow, R., and Inclen, W. "Leitstand fuer ein flexibles Fertigungssystem." Zeitschrift fuer wirtschaftliche Fertigung, Vol. 79, No. 5, pp. 219-222 (1984).
Gaimon, C. "Closed Versus Open Loop Dynamic Game Results on the Acquisition of New Technology." Working Paper, Academic Faculty of Management Sciences, College of Business, Ohio State University, Columbus,
OH (1988).
Garrett, S.E. "Strategy First: A Case in FMS Justification," in Proceedings of the Second ORSA/TIMS Conference on Flexible Manufacturing Systems (Ann Arbor, MI), K.E. Stecke and R. Suri (Eds.), Elsevier, Amsterdam, The Netherlands, pp. 17-30 (1986).
Gehner, K.R., "Software in Manufacturing" in Systems and Technologyfor Advanced Manufacturing. K.M. Gardiner
(Ed.), Society of Manufacturing Engineers, Dearborn, MI, pp. 79-86 (1983).
Gerwin, D., "Do's and Don'ts of Computerized Manufacturing." Harvard Business Review, Vol. 60, pp. 107-116
(March-April 1982).
Gerwin, D., "An Agenda for Research on the Flexibility of Manufacturing Processes." International Journal
of Operations and Production Management, Vol. 7, No. 1, pp. 38--49 (1987).
Gerwin, D. "Manufacturing Flexibility in the CAM Era" Business Horizons, Vol. 32, No. 1 pp. 78-84 (1989).
Gerwin, D. and Tarondeau, J.C. "La flexibilit~ dans les processus de production: le cas de l'automobile." Revue
Francaise de Gestion, Vol. 46, pp. 37-46 (June-July-August 1984).
Gerwin, D. and Tarondeau, J.C. "Consequences of Programmable Automation for French and American Automobile
Factories: An International Case Study." In Production Management: Methods and Studies, B. Lev (Ed.), NorthHolland, Amsterdam, The Netherlands, pp. 85-98 (1988).
Gerwin, D. and Tarondeau, J.C. "International Comparisons of Manufacturing Flexibility." In Managing International Manufacturing, K. Ferdows (Ed.), Elsevier, Amsterdam, The Netherlands, pp. 169-185 (1989).
Goldhar, J.D. and Jelinek, M. "Plan for Economics of Scope." Harvard Business Review, Vol. 61, pp. 141-148
(November-December 1983).
Graham, M.B.W. "A Tale of Two FMSs." In Managing Advanced Manufacturing Technology, C.A. Voss (Ed.),
IFS Publications, London, UK, pp. 45 356 (1986).



Graham, M.B.W. and Rosenthal, S.R. "Flexible Manufacturing Systems Require Flexible People" Human Systems
Management, Vol. 6, No. 3, pp. 211-222 (1986).
Gray, A.E., Seidmann, A., and Stecke, K.E. "Tool Management in Automated Manufacturing: Operational Issues
and Decision Problem" W.P. #590, Graduate School of Business Administration, University of Michigan, Ann
Arbor, MI (November 1988).
Groover, M.P. Automation, Production Systems and ComputerIntegrated Manufacturing. Prentice Hall, Englewood
Cliffs, NJ (1987).
Gupta, D., Buzacott, J., and Gerchak, Y. "Economic Analysis of Investment Decisions in Flexible Manufacturing Systems?' Working Paper, Department of Management Sciences, University of Waterloo, Waterloo, Ontario, Canada (1988).
Gustavsson, S.O. "Flexibility and Productivity in Complex Production Processes." International Journal of Production Research, Vol. 22, No. 5, pp. 801-808 (1984).
Hall, D. and Stecke, K. "Design Problems of Flexible Assembly" In Proceedings of the Second ORSA/TIMS
Conference on Flexible Manufacturing Systems (Ann Arbor, MI), K.E. Stecke and R. Suri (Eds.), Elsevier,
Amsterdam, The Netherlands, pp. 145-156 (1986).
Hart, A.G. Anticipations, Uncertainty and Dynamic Planning. New York, NY (1940).
Harvey, O.J., Hunt, D.E., and Schroder, H.M. Conceptual Systems and Personality Organization. Wiley, New
York, NY (1961).
Hayes, R.H. and Schmermer, R.W. "How Should You Organize Manufacturing?" Harvard Business Review, Vol. 56,
pp. 105-108 (January-February 1978).
Hayes, R.H. and Wheelwright, S.C. Restoring Our Competitive Edge: Competing through Manufacturing. Wiley,
New York, NY (1984).
He, H. and Pindyck, R.S. "Investment in Flexible Production Capacity." MIT-EL 89-001 WP, Center for Energy
Policy Research, MIT, Cambridge, MA (1989).
Hedrich, P. "Flexibilitaet in der Fertigungstechnik durch Computereinsatz?' CW- Publikationen Miinchen, Horst
Wildemann (Ed.), Munich, West Germany (1983).
Henry, C. "Investment Decisions Under Uncertainty: The Irreversibility Effect." American Economic Review.
Vol. 64, pp. 1006-1012 (1974).
Herrmann, P. and Pferdmenges, R. "Flexible Fertigungskonzepte?' Zeitschrift des VereinsDeutscher Ingenieure
fur Maschinenbau und Metallbearbeitung (VDI-Z), Vol. 122, No. 15/16, pp. 667-677 (August 1980).
Hollard, M. and Margirier, G. "Nouveaux pro~s de production et implications macro-~conomiques--Contribution
au debat sur la flexibilite"' Notes et Documents, Institut de recherche e~:onomiqueet de planification du d~veloppement (IREP), Grenoble, France (February 1986).
Hollard, M., Margirier, G., and Rosanvollon, A. "L'automatisation avanc~e de la production dans les activit~s
d'usinage. Degr~ de diffusion, camct~ristiquestechniques et socio-~conomiquesdes syst~mes automatise~ flexibles
d'usinage" Project Report, Institut de recherche ~conomique et de planification du d~veloppement (IREP),
Grenoble, France (1986).
Hutchinson, G.K. "Flexibility Is Key to Economic Feasibility of Automating Small Batch Manufacturing?' Industrial Engineering, Vol. 16, No. 6, pp. 77-86 (1984).
Hutchinson, G.K. and Holland, J. "The Economic Value of Flexible Automation?' Journal of Manufacturing
Systems, Vol. 1, No. 2, pp. 215-228 (1982).
Hutchinson, G.K. and Sinha, D. "Quantification of the Value of Flexibility." Journal of Manufacturing Systems,
Vol. 8, No. 1, pp. 4%56 (1989).
Ito, Y. "Evaluation of FMS: State of the Art Regarding How to Evaluate System Flexibility." Robotic & ComputerIntegrated Manufacturing, Vol. 3, No. 3, pp. 327-334 (1987).
Jacob, H. "Unsicherheit und Flexibilitaet--Zur Theorie der Planung bei Unsicherheit" Zeitschriftfuer Betriebswirtschaft, Vol. 44, pp. 299-326 (1974/1975).
Jaikumar, R. "Flexible Manufacturing Systems: A Managerial Perspective." WP #1-784-078, Harvard Business
School, Boston, MA (January 1984).
Jaikumar, R. "postindustrial Manufacturing" Harvard Business Review, Vol. 64, pp. 69-76 (November-December
Jones, R.A. and Ostroy, J.M. "Flexibility and Uncertainty." Rev/ewof Economic Studies, Vol. 51, pp. 13-32 (1984).
Jubin, M. "Etude exploratoire: Ateliers flexibles." Technical Report, Centre Technique des Industries M~caniques,
St. Etienne, France (1981).



Kaplan, R.S. "Must CIM be Justified by Faith Alone?" Harvard Business Review. Vol. 64, No. 2, pp. 87-95 (1986).
Karmarkar, U. and Kekre, S. "Manufacturing, Configuration Capacity and Mix Decisions Considering Operational Costs." Journal of Manufacturing Systems, Vol. 6, pp. 315-324 (1987).
Kearney & Trecker Corp. Manufacturing Systems Application Workbook. Managers Edition, Kearney and Trecker
Corp., Milwaukee, WI (1982).
Kegg, R.L. "Flexible Manufacturing Trends and Data Base Requirements." In Proceedings of the 5th International Conference on Production Engineering, ISPE, Tokyo (1984).
Kickert, W.J. "The Magic Word Flexibility." International Studies of Management and Organization, Vol. 14,
No. 4, pp. 6-31 (1985).
Klahorst, H.T. "Flexible Manufacturing Systems: Combining Elements to Lower Costs, Add Flexibility." Industrial Engineering, Vol. 13, No. 11, pp. 112-117 (November 1981).
Klein, B. and Meckling, W. '9,pplication of Operations Research to Development Decisions,' Operations Research,
Vol. 6, pp. 352-363 (May-June, 1958).
Knipschild, D. "Wirtschafflichkeit und Nutzen von fahrerlosen Transportsystemen" Zeitschriftfuer Wirtschafiliche
Fertigung, Vol. 81, No. 12, pp. 677-681 (1986).
Kolodny, H.E "Work Organization Design in Sweden: Some Impressions from 1982-83." Human Systems Management, Vol. 5, pp. 207-217 (1985).
Kolodny, H.E "Product Focussed Forms to Complement Flexible Technologies,' Working Paper, Faculty of Management, University of Toronto, Toronto, Canada (1989).
Koopmans, T. "On Flexibility of Future Preference,' In Human Judgements and Opportunity, M.W. Shelley and
G.L. Bryan (Eds.), Wiley, New York, NY, pp. 243-254 (1964).
Kozan, K., "Work Group Flexibility: Development and Construct Validation of a Measure." Human Relations,
Vol. 35, No. 3, pp. 239-258, 1982.
Krasa, A. and Llerena, P. "L'dvaluation de l'automatisation flexible dans les PMI: une comparaison francoallemande," Project Report, BETA, Universitd Louis Pasteur, Strasbourg, France (1987).
Kreps, D., 'A Representation Theorem for Preference for Flexibility." Econometrica, Vol. 47, pp. 565-577 (1979).
Kulatilaka, N. "Valuing the Flexibility of Flexible Manufacturing Systems." IEEE Transactions on Engineering
Management, Vol. 35, No. 4, pp. 250-257 (1988).
Kulatilaka, N. and Marks, S.E. "The Strategic Value of Flexibility: Reducing the Ability to Compromise,' American
Economic Review, Vol. 78, No. 3, pp. 574-580 (1988).
Kumar, V., "On Measurement of Flexibility in Flexible Manufacturing Systems: An Information-Theoretic Approach." In Proceedings of the Second ORSA/TIMS Conference on Flexible Manufacturing Systems (Ann Arbor, MI), K.E. Stecke and R. Suri (Eds.), Elsevier, Amsterdam, The Netherlands, pp. 131-143 (1986).
Kusiak, A. "Parts and Tools Handling Systems." In Modeling and Design of Flexible Manufacturing Systems,
A. Kusiak (Ed.), Elsevier, Amsterdam, The Netherlands, pp. 99-109 (1986).
Lam, A. "Measurement of a Flexible Manufacturing System." Term Project, University of Toronto, Toronto,
Canada (Winter 1988).
Lasserre, J.B. and Roubellat, F. "Measuring Decision Flexibility in Production Planning." 1EEE Transactions
on Automatic Control, Vol. AC-30, No. 5, pp. 447-452 (1985).
Lavington, E The English Capital Market, Methuen, London, UK (1921).
Leaver, E.W. and Brown, J.J. "Machines Without Men,' Fortune, pp. 165, 192,194, 196, 200, 203, 204 (November
Lederer, P. and Singhal, V. "Financial Justification of New Technologies,' Working Paper, Simon School of Business,
University of Rochester, Rochester, NY (1988).
Lim, S.H. "Flexibility in Flexible Manufacturing System--A Comparative Study of Three Systems." In Managing
Advanced Manufacturing Technology, C.A. Voss (Ed.), IFS Publications, London, UK, pp. 125-147 (1986).
Lim, S.H. "Flexible Manufacturing Systems and Manufacturing Flexibility in the United Kingdom,' International Journal of Operations and Production Management, Vol. 7, No. 6, pp. 44-54 (1987).
Lindholm, R. Job Reforms in Sweden, Swedish Employers Confederation, Stockholm, Sweden (1975).
McLean, C., Mitchell, M. and Barkmeyer, E. "A Computer Architecture for Small-Batch Manufacturing." IEEE
Spectrum, Vol. 20, pp. 59-64 (1983).
Maier, K. Die Flexibilitaet betrieblicher Leistungsprozesse, Harri Deutsch, Thun-Frankfurt/Main, West Germany



Maly, M. "Strategic, Organizational and Social Issues of CIM: International Comparative Analysis (Part I)."
WP-88-077, IIASA, A-2361 Laxenburg, Austria.
Mandelbaum, M. "Flexibility in Decision Making: An Exploration and Unification." Ph.D. Thesis, Department
of Industrial Engineering, University of Toronto, Toronto, Canada (1978).
Mandelbanm, M. and Buzacott, J. "Flexibility and Its Use: A Formal Decision Process and Manufacturing View."
In Proceedings of the Second ORSA/TIMS Conference on FMS (Ann Arbor, MI), K.E. Stecke and R. Suri
(Eds.), Elsevier, Amsterdam, The Netherlands, pp. 119-130 (1986).
March, J.G. and Simon, H.A. Organizations, Wiley, New York, NY (1958).
Marschak, T. and Nelson, R. "Flexibility, Uncertainty and Economic Theory." Metroeconomica, Vol. 14, pp.
42-58 (1962).
Mass6, P. Le choix des investissements--criteres et mbthodes, Dunod, Paris, France (1968).
Meffert, H. "Zum Problem der betriebswirtschaftlicben Flexibilitaet." Zeitschriftfuer Betriebswirtschaft, Vol.
39, pp. 779-800 (1969).
Melcher, A.J. and Booth, D. 'Advanced Manufacturing Systems: Characteristics and Strategic Implications."
Presented at ORSA/TIMS, St. Louis, MO (October 25-28, 1987).
Merkhofer, M.W. "The Value of Information Given Decision Flexibility." Management Science, Vol. 23, No.
7, pp. 716-727 (1977).
Milgrom, P. and Roberts, J. "Communication and Inventory as Substitutes in Organizing Production." Scandinavian Journal of Economics, Vol. 90, No. 3, pp. 275-289 (1988).
Milgrom, P. and Roberts, J. "The Economics of Modern Manufacturing: Technology, Strategy and Organization." Stanford University, CA (1989a).
Milgrom, P. and Roberts, J. "Rationalizability, Learning and Equilibrium in Games with Strategic Complimentarities." Discussion Paper, Graduate School of Business, Stanford University, CA (1989b).
Miller, S. "Industrial Robotics and Flexible Manufacturing Systems." In The Management of Productivity and
Technology in Manufacturing, P. Kleindorfer (Ed.), Plenum Press, New York, NY (1985).
Miller, B.L. "Scart's State Reduction Method, Flexibility, and A Dependent Demand Inventory Model." Operations Research, Vol. 34, No. 1, pp. 83-90 (1986).
Mintzberg, H. The Structure of Organizations, Prentice-Hall, Englewood Cliffs, NJ (1979).
Monahan, G. and Smunt, T. ' A Multi-level Decision Support System for Financial Justification of Automated
Manufacturing Systems." Interfaces, Vol 17, No. 6, pp. 29-40 (1987).
Monden, Y. "Toyota Production Smoothing Methods: Part 1." Industrial Engineering, Vol. 13, No. 8, pp. 42-51
(August 1981).
Myers, S. "Finance Theory and Financial Strategy." Interfaces, Vol. 14, No. 1, pp. 126-137 (1984).
Newman, W.E., Jr. "Models to Evaluate the Benefits of FMS Pallet Flexibility." In Proceedings of the Second
ORSA/TIMS Conference on Flexible Manufacturing Systems (Ann Arbor, MI), K.E. Stecke and R. Suri (Eds.),
Elsevier, Amsterdam, The Netherlands, pp. 209-220 (1986).
Ollus, M. and Mieskonen, J. "Bases for Flexibility in a Small Company: Some Issues of the Finnish TESProgramme." In Trends and Impacts of Computer Integrated Manufacturing. J. Ranta (Ed.), IIASA, A-2361
Laxenburg, Austria, pp. 355-373 (1989).
Panzar, J.C. and Wlllig, R.D. "Economics of Scope." American Economic Review, Vol. 71, No. 2, pp. 268-272 (1981).
Peter, M. "Vorgehensweise zur systematiscben Konzeption und Wirtschaftlichkeitsbeurteilung flexibler Fertigungssysteme." Diplomarbeit, Universitaet Karlsruhe, Karlsrube, West Germany (1984).
Porteus, E. "Investing in Reduced Setups in the EOQ Model." Management Science, Vol. 31, pp. 998-1010 (1985).
Preece, D.A. "Organizations, Flexibility and New Technology." In Managing Advanced Manufacturing Technology,
C.A. Voss (Ed.), IFS Publications, London, UK, pp. 367-382 (1986).
Primrose, P.L. and Leonard, R. "Conditions under which Flexible Manufacturing is Financially Viable." In Proceedings of the 3rd International Conference on FMS, IFS (Publications) Ltd., Bedford, U.K. (1984).
Proth, J-M. "La flexibilit6 des syst~mes de production: les 6volutions possibles." Revue Francaise de Gestion,
Vol. 35, pp. 12-20 (March, April, May 1982).
Pugh, D.S., Hickson, D.J., Hinings, C.R., and Turner, C. "Dimensions of Organizational Structure." Administrative
Science Quarterly, Vol. 13, No. 1, pp. 65-105 (1968).
Ranky, P. Computer-Integrated Manufacturing, Prentice-Hall, Englewood Cliffs, NJ (1981).



Ranta, J. "The Impact of Electronics and Information Technology on the Future Trends and Applications of CIM
Technologies?' In Trends and Impacts of Computer-Integrated Manufacturing, J. Ranta (Ed.), IIASA, A-2361
Laxenburg, Austria (1989).
Ranta, J. and Tchijov, I. "Economic and Success Factors of Flexible Manufacturing Systems: The Conventional
Explanation Revisited?' InternationalJournal of FlexibleManufacturingSystems, Vol. 2, No. 3, pp. 169-190 (1990).
Ranta, J. and Alabian, A. "Interactive Analysis of FMS Productivity and Flexibility." W.P.-88-098, IIASA, A-2361
Laxenburg, Austria (1988).
Rattner, L., Orne, D., and Wallace, W.A. "What Can We Learn from the Past?: A Comparison of Historical
Manufacturing Systems with Computer Integrated Manufacturing." Technical Report, Rensselaer Polytechnic
Institute, Troy, NY, 1988.
Reich, O.D. "Maintaining a Flexible Organization for Changing Conditions." American Management Assoc. New
York, NY (1932).
Rempp, H. "Einsatz flexibler Fertigungssysteme--Technische, einfuehrungsorganisatorische, wirtschaftliche und
arbeitsplatzbezogene Aspekte?' Werkstan und Betrieb, Vol. 115, No. 3, pp. 175-182 (1982).
Richard, A. "Quelques applications financi~res de la valeur d'option: structure des taux d'int~r& et actifs conditionnels?' In Flexibilitbs, information et d~cisions, P. Cohendet and P. Llerena (Eds.), Economica, Paris, France,
pp. 199-225 (1989).
Riebel, P. Die Elastizitaet des Betriebes, Westdeutscher Verlag, Koeln-Opladen, West Germany (1954).
Roeller, L-H. and Tombak, M. "Competition and Investment in Flexible Technologies?' Working Paper, INSEAD,
Fountainebleau Cedex, France (1989).
Rokeach, M. The Open and Closed Mind, Basic Books, New York, NY (1960).
Ropohl, G. "Zum Begriff der Flexibilitaet." Werkstattstechnik Vol. 57, pp. 644 (1967).
Ropohl, G. Flexible Fertigungssysteme, Krausskopf Verlag, Mainz, West Germany (1971).
Rosenhead, J., Elton, M., and Gupta, S.K. "Robustness and Optimality as Criteria for Strategic Decisions?'
Operational Research Quarterly, Vol. 23, pp. 413-431 (1972).
Scharf, P. "Strukturalternativen integrierter flexibler Fertigungsysteme und ihre Bewertung." Ph.D. dissertation,
Zugl., Stuttgart Univ., Stuttgart, West Germany (1975).
Schonberger, R.J. Japanese Manufacturing Techniques, The Free Press, New York, NY (1982).
Sethi, S. and Sorger, G. "Concepts of Forecast Horizons in Stochastic Dynamic Games?' Proceedings of 28th
1EEE CDC Conference, Tampa, Florida (December 1989).
Shaw, M.J. and Whinston, A.B. "A Distributed Knowledge-Based Approach to Flexible Automation: The Contract
Net Framework." International Journal of Flexible Manufacturing Systems, Vol. 1, No. 1, pp. 85-104 (1988).
Simon, H.A. The New Science of Management Decision, Revised Edition, Prentice-Hall, Englewood Cliffs, NJ
Slack, N. "The Flexibility of Manufacturing Systems." InternationalJournal of Operations and ProductionManagement, Vol. 7, No. 4, pp. 35-45 (1987).
Sorge, A., Hartmann, G., Warner, M., and Nicholas, I. Microelectronicsand Manpower in Manufacturing, Gower
Press, Aldershot, UK 1983.
Son, Y.K. and Park, C.S. "Economic Measure of Productivity, Quality and Flexibility in Advanced Manufacturing Systems." Journal of Manufacturing Systems, Vol. 6, No. 3, pp. 193-207 (1987).
Stecke, K. "Algorithms for Efficient Planning and Operation of a Particular FMS?' International Journal of Flexible Manufacturing Systems, Vol. 1, pp. 287-324 (1989).
Stecke, K.E. and Browne, J. "Variations in Flexible Manufacturing Systems According to the Relevant Types
of Automated Materials Handling." Material Flow, Vol. 2, pp. 179-185 (1985).
Stecke, K. and Raman, N. "Production Flexibilities and Their Impact on Manufacturing Strategy." W.E #484,
Graduate School of Business Administration, The University of Michigan, Ann Arbor, MI (December 1986).
Stecke, K. and Kim, I. "Performance Evaluation for Systems of Pooled Machines of Unequal Sizes: Unbalancing versus Balancing." European Journal of Operational Research, Vol. 42, No. 1, pp. 22-38 (1989).
Suresh, N. and Meredith, J. "Justifying Multi-machine Systems: An Integrated Strategic Approach." Journal
of Manufacturing Systems, Vol. 4, No. 2, pp. 117-134 (1986).
Swamidass, P.M. Manufacturing Flexibility, Monograph No. 2, Operations Management Association, Waco, Texas
Sweeney, M.T. "Flexible Manufacturing Systems--Managing Their Integration.'" In Managing Advanced Ma/tufilcturing Technology, C.A. Voss (Ed.), IFS Publications, London, UK, pp. 69-80 (1986).



Swoboda, E Die betriebliche Anpassung als Problem des betrieblichen Rechnungswesen, Dr. Gabler, Wiesbaden,
West Germany (1964).
Tombak, M. and De Meyer, A. "Flexibility and FMS: An Empirical Appraisal." IEEE Transactions on Engineering
Management, Vol. 35, pp. 101-107 (1988).
Tombak, M.M. "The Importance of Flexibility in Manufacturing." Working Paper, INSEAD, Fontainebleau,
France (1988).
Talaysum, A.T., Hassan, Z., Wisnosky, D., and Goldhar, J.D. "Scale vs. Scope: The Long-Run Economics of
the CIM/FMS Factory." In Advances in Production Management Systems, Vol. 85, E. Szelke and J. Browne
(Eds.), Elsevier, Amsterdam, The Netherlands (1986).
Tarondeau, J.C. "Produits et technologies--choix politique de l'entreprise industrielle." Dalloz, Paris, France (1982).
Tarondeau, J.C. "Technologies flexibles et performances." Working Paper, ESSEC, Cergy, France (1986).
Tintner, G. "The Theory of Choice under Subjective Risk and Uncertainty." Econometrica, Vol. 9, pp. 298-304
Triantis, A.J., and Hodder, J.E. "Valuing Production Flexibility When Demand Curves Are Downward Sloping." Working Paper, University of Wisconsin, Madison, WI (1989).
Vander Veen, D. and Jordan, WC. 'Analyzing Trade-Offs Between Machine Investment and Utilization." Operating
Sciences Department, General Motors Research Laboratory, Warren, MI (1988).
Wallace, W.J. and Thuesen, G.J. "Annotated Bibliography on Investment in Flexible Automation." The Engineering Economist, Vol. 32, No. 3, pp. 247-257 (1987).
Walton, R.E. "Establishing and Maintaining High Commitment Work Systems." In The Organization Life Cycle,
J.R. Kimberly, J.R. Miles, and Associates (Eds.), Jossey-Bass Publishers, San Francisco, CA (1980).
Warnecke, H.J. and Steinhilper, R. "Flexible Manufacturing Systems, EDP-Supported Planning; Application
Examples." In Proceedings of the First International Conference on Flexible Manufacturing Systems, IFS (Publ.)
Ltd., Kempston, Bedford, UK, pp. 345-356 (1982).
Wheelwright, S. "Japan--Where Operations Really Are Strategic.'" Harvard Business Review, Vol. 59, pp. 67-74
(July-August 1981).
Wildemann, H. Investitionsentscheidungsprozess fuer numerisch gesteuerte Fertigungssysteme, Dr. Gabler,
Wiesbaden, West Germany (1977).
Yao, D.D. "Material and Information Flows in Flexible Manufacturing Systems." Material Flow, Vol. 2, pp. 143-149
Yao, D.D. and Pei, EE "Flexible Parts Routing in Manufacturing Systems." Working Paper, Columbia University, New York, NY (September 1987).
Yilmaz, O.S. and Davis, R.E "Flexible Manufacturing Systems: Characteristics and Assessment." Engineering
Management International, Vol. 4, pp. 209-221 (1987).
Young, A.R. and Murray, J. "Performance and Evaluation of FMS." In Managing Advanced Manufacturing
Technology, C.A. Voss (Ed.), IFS Publications, London, UK, pp. 291-304 (1986).
Vives, X. "Technological Competition, Uncertainty and Oligopoly." Journal of Economic Theory, Vol. 48, pp.
386-415 (1989).
Yamashina, H., Okamura, K., and Matsumoto, K. "Flexible Manufacturing Systems in Japan--An Overview."
In Proceedings of the Fifth International Conference on Flexible Manufacturing Systems, IFS (Conferences)
Ltd., London, UK, pp. 405-416 (November 1986).
Zelenovic, D.M. "Flexibility--A Condition for Effective Production Systems." International Journal of Product
Research, IJ Productions, Vol. 20, No. 3, pp. 319-337 (1982).
Zhou, C. and Wysk, R.A. "Tool Management for Computer Integrated Manufacturing." In Proceedings of the
International Industrial Engineering Conference (Toronto, Canada), Institute of Industrial Engineers, Atlanta,
GA, pp. 603-607 (1989).