Decision Support

Systems: Issues and

Perspectives
- MICHAEL J. GINZBERG AND EDWARD A. STOHR

Objectives of the paper

What is DSS and how does it differ from the other computer-based
systems organizations have been using for years?

Examining definitions suggested about DSS

Issues highlighted and ignored by each definition

Attempt to outline the areas where further research could be more

fruitful

DSS Definitions

Support managerial decision makers in unstructured or semi-structured

decision situations ( Gorry and Scott Morton)

Extend his capabilities but not to replace his judgement

Model-based set of procedures for processing data and judgements to

assist a manager in his decision-making ( John Little)

DSS along five dimensions (Alter)

Use : active(DSS) vs. passive(EDP)

User : line, staff & management(DSS) vs. clerk(EDP)

Goal : overall effectiveness(DSS) vs. mechanical efficiency(EDP)

Time horizon : present & future(DSS) vs. past(EDP)

Objective : flexibility(DSS) vs. consistency(EDP)

DSS Definitions contd.

DSS consists of three interacting components ( Bonczek, Holsapple, &

Whinston)

A Language System, a Knowledge System, and a Problem Processing System

DSS as a product of development process in which the DSS user, the

DSS builder, and DSS itself are all capable of influencing one another
and resulting in evolution of the system

What DSS does to how the DSSs objective can be accomplished.

Concepts Underlying DSS Definitions

Central Issue in DSS

Support of decision making

Dimensions which attempt to narrow the field, to focus research along

some other dimension are missing the point

DSS is a computer-based information system used to support decision

making activities in situations where it is not possible or not desirable
to have an automated system perform the entire decision process

Anatomy : DSS technology

The elements which make up to DSS

Two areas are vital to the success of DSS concept

Technological progress determines what can be done

Behavioral research what should be done and how best to apply

technology to serve organizational goals

DSS taxonomy (Alter)

Three of the types are data-oriented

Four of the types are model-oriented

Some are equally oriented to both data retrieval and modelling while
others are graphics-oriented

DSS Software Classification Schemes

DSS Software Classification Schemes

- contd.

Institutional DSS deal with decisions of recurring nature

Portfolio Management System support investment managers

Ad hoc DSS deal with specific problems neither anticipated nor

recurring

Generalized Management Information System

Similar differences between Specific DSS and DSS generators

Procedural languages how data is to be retrieved or a computation

performed

Non-procedural languages what is required

Major Components of DSS

generators

Data Extraction System

Load the DSS data base with external data and data generated
internally to the organization by the MIS system

In model-oriented DSS

Management of both inputs and outputs of the models

Storage and access of the models themselves

Storage and access of a knowledge base of meta data concerning the

semantics of the model

DBMS issue

Use of time series and cross sectional data in modelling

Store intermediate and final results in DSS

Model Management System

Development of models and their subsequent use during sensitivity

analysis

Unresolved research issues

How can the results of sensitivity analysis be stored in the database and
related to underlying model assumptions and data values?

Use of artificial intelligence techniques

Semantic inheritance nets for knowledge representation

Resolution principle of first-order predicate calculus

Language Interface Subsystem

Compilers and interpreters that translate the statements and

commands used for specifying retrieval requests and defining models

How much non-procedurality can or should be attained with respect to

modelling languages

Concerns in the use of quasi-natural English like languages in DSS

User Interface Subsystem & System

Directory

Separate layer of software lying between the end-user and other

components of the DSS

Fertile area for DSS research

How should the user interaction be managed?

When should graphical displays be used?

System Directory gives the system some degree of self-knowledge

help user to learn the system

Physiology : DSS Usage Patterns

DSS must be used by managers or perhaps upper level managers not

a necessary condition for DSS

Portfolio Management System are used by investment decision makers

What DSS users have in common

All are decision makers or key stakeholders in the outcomes of the decision

Top managers are often decision ratifiers ratifies the decision made
by the staff

Degree of problem structure

Relative to particular decision makers

Physiology : DSS Usage Patterns

contd.

Semi-structured problem possible to bring some structure to bear on

the problem

Good DSS brings as much structure as the system user will accept

DSS supports all the phases

Intelligence or problem definition

Design of alternative solutions

Evaluation of and choice among the alternatives

Monitoring and control of implementation of the chosen solution

Type of support vary from phase to phase since the nature of activities
varies

Physiology : DSS Usage Patterns

contd.

DSS as appropriate to future oriented planning problems, not to current

control problems distinction hard to accept

Evolutionary stages as hallmark of DSS usage pattern for all systems,

not just DSS, evolve.

Characteristics of DSS usage patterns

Voluntary usage, interactive usage and unplanned usage

Controllability as against interactive

Two way communication as against one way communication

DSS Usage Patterns

Ontogeny : DSS Development

Patterns

Common prescription for DSS development

Normative decision modelling

Participative

Evolutionary

Normative modelling additional structure is brought to unstructured

decision situations

User involvement in DSS is essential

Problem or problem view changes and hence the system must evolve
to remain relevant and useful

Ontogeny : DSS Development

Patterns contd.

Evolutionary design in case of substantial uncertainty and

Prototyping in case of little uncertainty

DSS oriented towards particular individual needs and preferences of

that person

DSS oriented towards specific roles less attention to individual user

preferences

Organizational changes for successful DSS implementation

DSS development patterns - diagram

Thank You !