404 PAKT IV. GENERIC MODELING APP
works” Jounal ofthe Association for Compuut- Spearman, ML. D. L Woodru x
sng Machinery, 2"(2), 313-322 ‘ONWIP: Pull Alternative wo K
Schweitzer, Paul J. Abraham Seidman, and Sarit rernational Journal of Production Ri
Shalev-Oren (1986), “The Correction Tern 8, 879-804,
Research Letter an: A’ Precursor 10 Simulation fo
facturing Systems: A Comparative Evaluation stitute of Electrical and Electronic
1985), “Analysis of Flexible Manufacturing Sys- ing Flexible Manufacturing Systems Ui
Analysis of Manufacturing Systems,” Mamgac- Whit Ward [1983 veweing
Solberg, James J. [1980], “A Users Guide to CAN: 2755-2815,
PROBLEMS
11.1. The queueing models in this chapter pro: L144, Orders arrive ata 8 rly at
ice rates are constant through 1 t time, and mean an
assumptions would not apply: 11.8. An assemb rece
the shipment is late and a partal onder be- sk $ exponentially distributed wi
ng shipped. Most orders are small, requiring like 1-8 minutes. Find the mean and varCHAPTER 11 GENERAL
thus, in the long term the shop schedule averages
44 hours per week. The foreman has proposed
siving priority o a key customer who has been
‘complaining about long lead times. This customer
accounts for the orders. Estimate the
percent
effect on throughput time for the key customers
and others ifthe
ny.
stations. Order arrivals are a P
A flexible assembly sistem has five work:
mean rate of 18 per hour. Service time is expo:
‘ential with a rate of 4 per ‘workstation
Workstations place the
single mater
handler
1 Determine the
ting is Fan
m, find che interarrival time distribution at
b. Suppose the unloader s
Jers ti
two packing stations. If
ach packing station,
. Suppose the unloader
packing stat every second order goes
the same packer). What is the interarrival
distribution for packers?
118. Resolve Example 11.6 with py) = 0.25 and
an external arival rate of A= 8
(Qt. Three workers “kit” orders by pulling the
lied number of parts from a warehouse an
icing them in a toe
In fact, the computer scheduling sys
‘of kitting orders
in queue at the kitters. Time
>it an
utes Kits then go 10 an assembl
area. Four assem
available. Assemb
spected. Two inspectors are available. Inspection
lime is exponer 15 minutes
minutes. Assembled kits ate then in
ial with mear Find
the average number of ord
station and the average time for
process at exch
11.10. ody Problem 11.9 such that
n with probability
kit fils
nspeet 4. Failed kits are re
ANUFACTURING SYSTEMS. ANALYTICAL QUEUEING MODELS
405
turned to assembly: Rework time is als
zes and throughput time
OTE in tubs and are transported on pallets. Eat
pallet holds one batch of parts. all parts are moved
between machines by forklift The shop produces
four types of pars. Their routings are shown in
Table 11.6 Find the effective arrival rates at each
workstation and the
maz,
rile
res output from an MAE process,
r
three part pes are 10, 5, and 20 per day, respec
tively Service time
par pes are processed. Arrival rates for the
days for pant ype 2, and deterministic at 001 days
for part type 3. If part type 1 has highest priority
and part type 2 has higher priority than part type
3, find the expected number of each part ype and
throughput times atthe workstation
11.13. Show that the balance equation 11.34 in
Example 11.11 is satisfied bythe state probabilities
fined! in equation 11.35, (Hint: Find the v
counts and service rates for workstations, Then re
place the p
11.14, Reconsider Example 116.
that have been returned to assembly bec
failed inspection the first time need not be rein
After a second visit to assembly, these jobs
babi
Suppose jobs
we system. Find the new transfer pn
and resolve
1 examp!
11.15. Discuss the diferent assumptions of open
and closed queueing newwork models, Contronted
‘ith a real system, what factors would determine
‘which model
11.16. Three prospective designs are under con
sideration for a manufacturing cell The first uses ¢
F ate 4, com:
mon queue, and total arrival rate A. The second
‘option uses the same machines but has a sepa
rate queue for each machine. Arriving part bate!
a
‘Table 11.6 Part Routings and Demand for Problem 11.11
Par Type Opt Op.2
H MB
3 Ms
ms Ops Batches/Period
Mi 100406 ART IV. GENERIC MODELING APPROACHES
will be randomly assigned to a machine The third 11.20, Resolve Problem 11.198 assuming that am
uration consists of one fast machine, which ple material handlers exists so that parts never
serves atthe rate 34. Ignoring capital and operat- need to wait for material handling
oss, which system has the lowest throughput 43.24, Resolve Problem 11.196
time and WIP levels? You may assume interat 7 ckly ¢
work Arrival rates are given by weekly demand.
and service times are exponential and let x = 3
dition to finding average throughput time for
and A= 10,
an arbitrary job, find the average throughput tin
ALA7. A job shop has three types of machines; for product type 1 specifically
two mills, one drill press, and one surface grinder: 49,22, Re-solve Problem 11.19% assuming that wo
About 60 percent ofthese go to m The
40 hours Use the throughs time eqution 11.25
other 40 percent start at the drill, One half of the '
dein ob go nex vo milling where he oer 23) ourmachine production ser (al
sap hig ee routing probabilities are as shown in the follow
ers leave the shop Job aeays leave the stem 86
ier grinding Operation times are exponentially IN ‘bles. Seven jobs are Kept in process at all
Jistibuted, averaging one day per job for milling, MES Service times are exponentially distribute
drilling, and grinding, Find the average num!
jobs in the sytem
11.8, Resolve problem 11.17 assuming that op
tration times are uniformly distributed between 0
7 Time (br
and 2 days per jot
1149. Using the closed network of que
‘model, analyze the following problem. A job shop
nas three machines. Four products are made ac .
elie oR ig Ge eee ee Intermachine Routing Probabilities
lable, and it takes him about 5 minutes t0 a
move a load of material berween machines. Yo
may assume that the handler is aot needed to
posit new parts to the first operation, Machines
and 2 run 40 hours per week, and machine 3
runs SD ours. Assume that we try to keep one
week's demand worth of work in process at
1. Analyze by explicitly modeling the four sep- a. Find the relative visit counts and utlizations
Db. Aggregate nto a generic, Solve this network using MVA Find the pro
‘weighted-verage) product and reanaly duction rate
Table 11.7 Product Routings for Problem 11.19
Machine Time in Hi Weekly Demand
1 0 1