See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/241802417

Flow Shop Scheduling Problem: a Computational Study

Article · January 2002

CITATIONS READS
5 832

3 authors, including:

Amr Arisha Paul Young

TU Dublin Dublin City University
86 PUBLICATIONS 1,847 CITATIONS 74 PUBLICATIONS 861 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Amr Arisha on 12 February 2014.

The user has requested enhancement of the downloaded file.

 Dublin Institute of Technology
ARROW@DIT
Conference papers School of Marketing

2002-01-01

Flow Shop Scheduling Problem: a Computational

Study
Amr Arisha
Dublin Institute of Technology,, amr.arisha@dit.ie

Paul Young
Dublin City University

Mohie El Baradie
Dublin City University

Follow this and additional works at: http://arrow.dit.ie/buschmarcon

Part of the Other Operations Research, Systems Engineering and Industrial Engineering
Commons

Recommended Citation
Arish, A., Young, P., El Baradie, M.:Flow Shop Scheduling Problem: a Computational Study.Sixth International Conference on
Production Engineering and Design for Development (PEDD6),Cairo, Egypt, pp 543 – 557.

This Conference Paper is brought to you for free and open access by the
School of Marketing at ARROW@DIT. It has been accepted for inclusion
in Conference papers by an authorized administrator of ARROW@DIT.
For more information, please contact yvonne.desmond@dit.ie,
arrow.admin@dit.ie.

This work is licensed under a Creative Commons Attribution-

Noncommercial-Share Alike 3.0 License
FLOW SHOP SCHEDULING PROBLEM: A COMPUTATIONAL STUDY

A. Arisha, P. Young, and M. EL Baradie

School of Mechanical & Manufacturing Engineering, Dublin City University
Dublin 9, Ireland
Email: amr.arisha@gmail.com

ABSTRACT:

A computational study has been developed to obtain optimal / near optimal solution for the flow
shop scheduling problem with make-span minimization as the primary criterion and the
minimization of either the mean completion time, total waiting time or total idle time as the
secondary criterion. The objective is to determine a sequence of operations in which to process ‘n’
jobs on ‘m’ machines in same order (flow shop environment) where skipping is allowed. The
Simulation approach for deterministic and stochastic flow shop scheduling has been developed. It
reads and manipulates data for 500 jobs on 500 machines. Different factorial experiments present a
comparative study on the performance of different dispatching rules, such as FCFS, SPT, LPT,
SRPT and LRPT with respect to the objectives of minimizing makespan, mean flow time, waiting
time of jobs, and idle time of machines.
The proposed model is evaluated and found to be relatively more effective in finding optimal/ near
optimal solutions in many cases. The influence of the problem size in computational time for this
model is discussed and recommendations for further research are presented.

KEYWORDS: Flow Shop Scheduling, Simulation, Dispatching Rules, and Enumerative

Optimization.

1. INTRODUCTION

The purpose of this paper is twofold: (1) to provide a simulation program able to find the optimum /
near optimum sequence for general flow shop scheduling problem with make-span minimization as
main criteria; (2) to compare different dispatching rules on minimizing multiple criteria.
Numerous combinatorial optimization procedures have been proposed for solving the general
flowshop problem with the maximum flow time criterion. Many researches have been successful in
developing efficient solution algorithms for flowshop scheduling and sequencing [1, 2, 3, 4, 5, 6, 7
and 8] using up to 10 machines. Dannenbring [2] found that for small size shop problems his
heuristic outperformed others in minimizing the make-span for 1280 flowshop scheduling
problems. Ezat and El Baradie carried a simulation study for pure flowshop scheduling with make-
span minimization as a major criterion for n ψ90 on m ψ90 [9]. In This paper study general flow
shop scheduling problem with make-span minimization as main criteria for n ψ 250 and m ψ 250
with different ranges of random numbers generated (0-99) for processing times matrix.
2. THE FLOWSHOP SCHEDULING PROBLEM

The flowshop problem has interested researchers for nearly half a century. The flowshop problem
consists of two major elements: (1) a production system of ‘m’ machines; and (2) a set of ‘n’ jobs to
be processed on these machines. All ‘n’ jobs are so similar that they have essential the same order
of processing on the M machines, Fig. 1. The focus of this problem is to sequence or order the ‘n’
jobs through the ‘m’ machine(s) production system so that some measure of production cost is
minimized [10]. Indeed, flowshop scheduling problem has been shown to be NP-complete for non-
preemptive schedules [11].
Input Input Input Input

M/c #1 M/c #2 M/c #3 M/c #m

Output Output Output Output

Fig.1: Work flow in General Flow Shop Scheduling Model

The assumptions of the flowshop problem are well documented in the production research literature
[3,4,5,18]. In summary:
1) All ‘n’ jobs are available for processing, beginning on machine1, at time zero.
2) Once started into the process, one job may not pass another, but must remain in the same
sequence position for its entire processing through the ‘m’ machines.
3) Each job may be processed on only a single machine at one time, so that job splitting is not
permitted.
4) There is only one of each type of machine available.
5) At most, only one job at a time can be processed on an individual machine.
6) The processing times of all ‘n’ jobs on each of the ‘m’ machines are predetermined.
7) The set-up times for the jobs are sequence independent so that set-up times can be considered
a part of the processing times.
8) In-process inventory is allowed between consecutive machines in the production system.
9) Non-preemption; whereas operations can not be interrupted and each machine can handle
only one job at a time.
10) Skipping is allowed in this model.

3. THE PERFORMANCE CRITERIA

The performance criteria are those most commonly used as proposed by Stafford [15], for
optimizing the general flowshop model.

1. Makespan
Throughout the half century of flowshop scheduling research, the predominant objective
function has been to minimize make-span. [10]
The expression used is as follows:
Minimize: Cmax
 2. Mean Completion Time
Conway et al. (1967), Panwalker and Khan (1975), Bensal (1977), and Scwarc (1983) have
all discussed mean job completion time or mean flow time as an appropriate measure of the
quality of a flowshop scheduling problem solution. Mean job completion time may be
expressed as follows:
n
C = ∑ Job completion times / n
i =1
3. Total Waiting Time
Minimizing total job idle time, while the jobs wait for the next machine in the processing
sequence to be ready to process them, may be expressed as follows:
m n
W( nxm ) = ∑∑ wij
i =1 j =1

4. Total Idle Time

Overall all machine idle time will be considered in this model (the time that machines 2,…. ,
M spend waiting for the first job in the sequence to arrive will be counted). Overall machine
idle time may be minimized according to the following expression:
m n
Minimize: ∑∑ X
1=1 j =1
ij

4. DISPATCHING RULES

A dispatching rule is used to select the next job to be processed from a set of jobs awaiting service
at a facility that becomes free. The difficulty of the choice of a dispatching rule arises from the fact
that there are n! ways of sequencing ‘n’ jobs waiting in the queue at a particular facility and the
shop floor conditions elsewhere in the shop may influence the optimal sequence of jobs at the
present facility [12].
Five basic dispatching rules have been selected to be investigated in this research. A brief
description about each rule will be presented:

Rule (1) FCFS (First Come First Served): This rule dispatches jobs based on their arrival times
or release dates. The job that has been waiting in queue the longest is selected. The FCFS rule
is simple to implement and has a number of noteworthy properties. For example, if the
processing times of the jobs are random variables from the same distribution, then the FCFS
rule minimizes the variance of the average waiting time. This rule tends to construct schedules
that exhibit a low variance in the average total time spent by the jobs in this shop.

Rule (2) SPT (Shortest Processing Time): The SPT first rule is a widely used dispatching rule.
The SPT rule minimizes the sum of the completion times ΣCj (usually referred as the flow
time), the number of jobs in the system at any point in time, and the average number of jobs in
the system over time for the following machine environments: set of unique machines in series,
the bank of identical machines in parallel, and the proportionate flow shop.

Rule (3) LPT (Longest Processing Time): The LPT rule is particularly useful in the case of a
bank of parallel machines where the make-span has to be minimized. This rule selects the job
with the longest processing (from the queue of jobs) to go next when a machine becomes
available. Inherently, the LPT rule has a load balancing property, as it tends to avoid the
situation where one long job is in process while all other machines are free. Therefore, after
using the LPT rule to partition the jobs among the machines, it is possible to resequence the
 jobs for the individual machines to optimize another objective besides make-span. This rule is
more effective when preemption is allowed.

Rule (4) SRPT (Shortest Remaining Processing Time): The SRPT is a variation of SPT that is
applicable when the jobs have different release dates. SRPT rule selects operations that belong
to the job with the smallest total processing time remaining. It can be effective in minimizing
the make-span when preemption is allowed.

Rule (5) LRPT (longest Remaining Processing Time): The LRPT is a variation of LPT that
selects the operations that belong to the job with the largest total processing time remaining.
LRPT rule is of importance when preemption is allowed and especially in parallel identical
machines. LRPT rule always minimizes the idle time of machines.

5. EXPERIMENTAL DESIGN

In flowshop sequencing research, the standard approach for evaluating a new problem solving
technique, whether it is a heuristic or an optimization model, is to generate a set of problems of
different sizes, and then to solve this
common set of problems with the
new techniques and with one or Make-Span vs Different Number of Runs
more other proven techniques 80
designed for the same flowshop 75
problem. 70

A computer simulation program 65

Table. 1 has been developed into two 60

FCFS
MakeSpan

SPT
phases: (1) to find the optimum/near 55 LPT
SRPT
optimum solution for general 50
LRPT

flowshop problem to minimize the 45

makespan; (2) to measure the 40

35
effectiveness of various priority rules
30
for flow shop scheduling. The 1 10 30 100 300 500 900

program can read data up to 500 x No. of Runs

500 and use both deterministic and Fig.2: A Steady-State Analysis of the Model
stochastic processing time input. The
input processing times may be generated from different seed random numbers for each single run or
read directly data from an input file. The number of runs for each case is 300 where the results turn
to steady state started as shown in Fig. 2.
Different factorial sets of experiments were conducted to verify that the program would provide
optimal solutions to general flowshop problems and to compare between various dispatching rules.

6. EXPERIMENTAL RESULTS

Phase 1:

This phase has multi-objectives for n/m/F/Cmax problem. It can provide the followings:
1) All the job sequences and their correspondent makespan for each sequence.
2) The optimal job sequence and its makespan value.
3) Frequencies for all job sequences.
4) CPU time for the solution.
A sample of the output of the program and optimal makespan are shown in Fig. 3. Although, many
researchers have been working on the flowshop sequencing problem for many years, it has been
found nowhere any results about the distribution of the objective function and the distribution of the
optima of this function. In effect, such an approach gives an intuitive idea about the problem and is
important to allow the reader to judge the quality of the method used for this problem [13]. The
distribution of all the possible make-spans obtained by complete enumeration of two different
problems are given in Fig. 4. This distribution is given relative to the optimum solution. The
processing times were randomly generated (integers between 1 and 10). The distribution seems to
be almost symmetric and its range is contained in an interval of 20% around the mean. A χ2 test
does neither confirm nor refute that this distribution is Gaussian, therefore the use of the mean
makespan given by a heuristic seems to be meaningful.

Table 1: Terminology associated with Simulation Model

Model parameters Decision Variables

Model n/m/F/Cmax n > 0 i = 1, …….., n,

n number of jobs to be processed, n < 250 ( recommended ) ( the model
can read up to 500 jobs ),
m number of machines ( processing m > 0 j = 1, …….., m,
steps), m < 250 ( recommended) ( the model
F general flowshop scheduling problem, can read up to 500 machines ),
Number of runs,
Cmax the criterion is Make-Span,
Number of seed,
Pij processing time off job ‘j’ on machine
‘i’,
C the criterion is Mean Completion Time,
wij waiting time before start the job ‘j’,
Objective functions
Xij idle time of machine ‘i’ before start job
in position j in the sequence, 1) Make-span
W(nxm) total waiting time. Minimize: Cmax

2) Mean Completion Time

n m
Minimize: C = 1
n ∑∑ (w
j =1 i =1
ij + Pij )
Dispatching Rules
3) Total Waiting Time
m n
FCFS : First Come First Served Minimize: W(nxm) = ∑∑ w ij
SPT : Shortest Processing Time i =1 j =1

LPT : Longest Processing Time 4) Total Idle Time

SRPT : Shortest Remaining Processing Time m n

LRPT : Longest Remaining Processing Time Minimize: ∑∑ X

1=1 j =1
ij
 5 jobs on 7 machines

n\ m M1 M2 M3 M4 M5 M6 M7
J1 8 9 7 8 6 2 8
J2 3 5 9 7 5 2 4
J3 9 1 8 0 10 8 2 Input matrix of processing times
J4 3 8 10 2 3 0 0
J5 4 2 6 8 10 1 0

5 Jobs, 7 Machines, Random seed=48

Number of runs = 1

12345 Make-Span=75
Job sequences 12354 Make-Span=75
12435 Make-Span=75
24153 Make-Span=71
24513 Make-Span=72
24531 Make-Span=67
31524 Make-Span=66
34125 Make-Span=66 The entire possible job sequences
34152 Make-Span=64
41532 Make-Span=66
45312 Make-Span=61
45321 Make-Span=63
54132 Make-Span=66
51342 Make-Span=64
51432 Make-Span=66
. . . .
. . . . .
Number of Job Sequences = 120

53421 = Optimum Job Sequence, Optimal Make-span = 60

CPU time in seconds =0.211000

Make-Span's Frequencies:

For the Make-Span =60: the Frequency =3

For the Make-Span =61: the Frequency =3
For the Make-Span =62: the Frequency =1
For the Make-Span =63: the Frequency =3
For the Make-Span =64: the Frequency =15
For the Make-Span =66: the Frequency =15
For the Make-Span =67: the Frequency =20 The frequency of
For the Make-Span =68: the Frequency =6 each make-span
For the Make-Span =70: the Frequency =1
For the Make-Span =71: the Frequency =39
For the Make-Span =72: the Frequency =1
For the Make-Span =73: the Frequency =3
For the Make-Span =75: the Frequency =10
53421= Optimum Job Sequence, Optimal Make-Span = 60
Optimal Job Sequence

Fig. 3: Sample of program output for Phase 1

8 Jobs, 8 Machines, Random seed=33 10 Jobs, 30 Machines, Random seed=25

Number of runs = 1 Number of runs = 1
Number of Job Sequences = 40320 Number of Job Sequences = 3628800

6,5,4,3,8,7,1,2 = Optimum Job 8,5,10,7,6,4,9,3,2,1 = Optimum Job

Sequence, Optimal Make-span = 67 Sequence, Optimal Make-span = 109
CPU time in seconds =61.519000 CPU time in seconds =1168.229000

Frequency Distribution Frequency Distribution

3500 250

3000
200
Frequency ( x 1000)

2500
Frequency

150
2000

1500 100

1000
50
500

0 0
9

7
67
69

71

73
75

77

79
81

83

85
87

89

91

93

95

97

99

10

11

11

11

12

12

12

13

13

13

13

14

14

14

15

15

15
10

Make-Span Make-Span

Fig. 4: Frequency distribution of two different flowshop problems

Phase 2:

To compare the dispatching rules under identical conditions, the same pseudo random numbers
generated per run, and the same number of runs. A different factorial experiment for the selected
rules, 7 machines shop (5, 20, 50, 80, 130, 200, 250), and 9 levels of Work In Process (WIP)
(number of jobs in shop) equal to (5, 10, 30, 50, 80, 100, 150, 200, 250) were executed.
Computational times for the entire experiments are shown in Table. 4. Samples of machine shop
were presented for each performance criteria.

Average Make-Span Criterion

For small machine numbers, Fig. 5, there is a clear spread across the different rules with SPT
rule providing the best results and the LPT rule performing worst.
For larger machine numbers, Fig. 6, the LPT rule is still clearly the worst, however the other
rules show almost identical results. Nevertheless the SPT rule is the best performer overall.

Average Mean Completion Time Criterion

For this criterion, Fig. 7, the SRPT rule provides the best results. Again, the LPT rule performs
worst, sometimes rivaled by the LRPT rule.

Average Waiting Time Criterion

This criterion changes the order of the rules with LPT performing best and SPT performing
worst, Fig. 8. In addition increasing the number of jobs increases the spread between best and
worst results significantly.

Average Idle Time Criterion

As with the waiting time criterion, Fig. 9 shows that the spread of the performance increases
with job number. Here the LRPT rule is clearly the best while the LPT rule performs worst.

Avg. Make-Span vs Different Dispatching Rules (20 Machines/ 300 Runs)

1800

1600

1400

1200
Avg. Make-Span

FCFS
1000
SPT
LPT
800 SRPT
LRPT

600

400

200

0
5 10 30 50 80 100 150 200 250
No. of Jobs

Fig. 5: Effect of the dispatching rules vs. average makespan under different machine shop
 Avg. Make-Span vs Different Dispatching Rules(200 Machines/ 300 Runs)

4000

3500

3000

2500
FCFS

Avg. Make-Span
SPT
LPT
2000
SRPT
LRPT

1500

1000

500

0
5 10 30 50 80 100 150 200 250
No. of Jobs

Fig. 6: Effect of the dispatching rules vs. average makespan under different machine shop
Avg. Mean Completion Time vs Different Dispatching Rules (20 Machines/300 Runs)

1200

1000
Avg. Mean Completion Time

800

FCFS
SPT
600 LPT
SRPT
LRPT

400

200

0
5 10 30 50 80 100 150 200 250
No. of Jobs

Fig. 7: Effect of the dispatching rules vs. average mean completion time under different machine shop

Avg. Waiting Time vs Different Dispatching Rules (80 Machines/300 Runs)

180000

160000

140000

120000
Avg. Waiting Time

FCFS
100000
SPT
LPT
80000 SRPT
LRPT

60000

40000

20000

0
5 10 30 50 80 100 150 200 250
No. of Jobs

Fig. 8: Effect of the dispatching rules vs. average waiting time under different machine shop
 Avg. Idle Time vs Different Dispatching Rules (200 Machines / 300 Runs)

250000

200000

Avg. Idle Time 150000

FCFS
SPT
LPT
SRPT
LRPT
100000

50000

0
5 10 30 50 80 100 150 200 250
No. of Jobs

Fig. 9: Effect of the dispatching rules vs. average idle time under different machine shop

7. RESULTS ANALYSIS

In general, the quality of a technique’s solutions is measured in at least two dimensions: (1) how
close the solution comes to the optimal solution if it can be measured; and (2) how much computer
time is required to solve problems of a given size.
Due to wide differences in software, platform, problem size, experimental design and reporting, it is
very difficult to compare the performance of different techniques directly. To allow some
comparison to be made Table 2 shows the average percentage increase over optimum make-span
time as reported by each of the researchers for their algorithms. To enhance the comparison, the
right column indicates the relative limitations of each model.

Table 2: Comparison between some different studies

Avg. % of
Method increased of Limitations
Optimum
Palmer, 1965 10 % - 35 % Optimum achieved in 30% of cases
Small scale problems only
Campbell, 1970 5 % – 20 % Optimum is not guaranteed
Economical n < 8 ,
Dannenbring, 1977 5 % – 15 % Optimum achieved in 35% of cases
n < 6 , m < 10 only
Gupta, 1971 10 % - 20 % Optimum is not guaranteed
Al-Qattan, 1990 0 % - 15 % Optimum is not guaranteed
Ezat – El Baradie, 1993 0 % - 10 % Optimum is for n < 12 m < 60
Pure flowshop scheduling problems only
Max. size n < 90 , m < 90
Tsang – Stafford, 2001 0%-5% Optimum is guaranteed for n < 7 , m < 7
LEKIN, 1998 [14] 0 % - 10 % Optimum is not guaranteed
Max. size n < 10 , m < 18
Arisha – El Baradie, 2001 0% Optimum is guaranteed for n < 50 , m < 250
General flowshop scheduling problems only
Max. size n < 500 , m < 500
0 % - 10 % For n > 50, m > 250
The exponential increase in solution time with number of jobs is shown in Table 3 for the phase 1
simulation model used in this work. The length of time for computation means that the full search
cannot be economically used where the number of jobs exceeds 30. The effect of machine numbers
on time is clearly much less significant.
Table 3: CPU time (seconds) to find optimum makespan for different problem sizes
nxm 5 10 20 40 100 250
5 0.161 0.18 0.17 0.181 0.21 0.22
6 1.1 1.1 1.2 1.2 1.3 1.3
7 8.1 8.2 8.4 8.4 8.8 9.0
8 40.3 43.3 44.9 45.6 44.3 45.2
9 232.1 232.6 245.0 235.3 245.1 248.1
10 1158.0 1176.3 1256.4 1307.3 1354.2 1298.3
11 13115.3 13118.3 14234.7 13215.2 13684.3 13968.3
12 56025.3 56036.9 57231.3 56016.7 57863.3 58015.2
20 - 551369.4 - 543652.4 - -
30 - - - 2605248.0 - -

10000000

1000000
Table 4: The factorial experiment for phase 2
Log CPU Times (seconds)

100000 and the CPU times

5 Machines
10000 40 Machines
100 Machines
1000
Number of Machines
250 Machines
n x m 5 20 50 80 130 200 250
100 5 17.25 17.9 18.05 18.56 19.6 20.25 21.82
10 18.45 18.96 19.5 19.91 20.45 21.68 22.35
Number of Jobs

10 30 19.48 19.9 20.45 21.34 21.98 22.57 23.01

50 20.01 20.45 20.98 21.54 21.87 22.15 23.24
1 80 21.14 21.29 21.76 22.14 22.56 22.89 23.96
5 6 7 8 9 10 11 12 20 30 100 21.82 22.15 22.6 23.04 23.42 24.12 24.54
Number of Jobs 150 22.15 22.35 22.98 23.45 23.87 24.36 24.95
200 22.59 22.84 23.15 23.68 24.15 24.59 25.15
250 23.15 23.59 23.96 24.15 24.93 25.09 25.96
Fig.10: CPU Times vs. different Number of Jobs

8. DISCUSSION

Flowshop scheduling is one of the most critical activities for the production planner. Minimizing
the make-span, mean flow time, job waiting times and machine idle time are the major objectives to
reduce the processing costs. The flowshop scheduling problem is NP-hard and several studies have
been done to solve small size flowshop problems.
The parameters that affect the size of flowshop problems are ‘n’ and ‘m’. The problem size
complexity is based on these two parameters. The results of the proposed model indicated that ‘n’
has much stronger influence on computer solution time than ‘m’. Based on these studies and the
proven NP-completeness of the problem, it is clear that ‘n’ is a much more important determinant of
computer solution time required for the flowshop problem.
The simulation study has been carried out under several operational conditions. It found the
optimum/ near optimum solution in a reasonable computational time for most cases in a specific
range of problem sizes. It is recommended to use the model for the problems where number of jobs
is less than 30 and number of machines is less than 250 as it is not economical for larger scale.
Furthermore, a comparative study on the performance of various dispatching rules has been carried
out under different shop machine
and utilization levels. The model
runs for 300 iteration using the
16.0
random generator for the same
specific shop conditions. It has 14.0
13.8

been observed that no single rule

11.8
performs well for all important 12.0

criteria related to completion time,

Avg. Percentage Increase

10.0 9.7
9.4
waiting time and idle time. SPT
has performed the best to minimize 8.0

make-span under different 6.5

6.0
conditions, as is clearly evident
from Fig. 11. The SPT rule is quite 4.0

often used as benchmark since it is

found to be very effective in 2.0

minimizing make-span and also 0.0

mean tardiness under highly loaded FCFS SPT LPT
Dispatching Rules
SRPT LRPT

shop floor conditions [12]. SPT

show the worst for job waiting Fig. 11: Comparison of average normalized increase over
criterion. While LPT shows the optimum make-span for different dispatching rules
worst performance for make-span
criterion, it tends to be the best rule to minimize the job waiting time especially for high utilization
level (n > 80). For the average mean completion time (mean flow time) SRPT shows the best
performance for different levels of utilization. LRPT and LPT performed worst for average mean
completion time criterion. LRPT rule tends to dominate with respect to the machine idle time while
LPT and SRPT showed the worst performance depending on job numbers.
While never being the best or worst performer for any criterion the FCFS rule is effective in
minimizing the maximum flow time and the variance in flow time. Its consistent “mid-table”
performance allows its use as a benchmark.

9. CONCLUSIONS

Some recommended trends for flowshop scheduling problem are listed down:

- A full solution search is uneconomic for larger sized problems so it is important to

develop intelligent search techniques that truncate the search tree size to get the optimum /
near optimum solution.
- Although hundreds of publications and studies had been done in flowshop scheduling
problems, but the need to investigate problem size beyond the small sizes (n < 9, m < 9) is
still required [10].
- Artificial Intelligent techniques turns to be one of the most effective tools to handle
optimization problem to get a satisfying solution [17].
- However, there are still many areas where more research is needed, including the integration
of the three common approaches; operations research-based, simulation-based and AI-based
in order to develop a comprehensive hybrid model to solve flowshop scheduling problems.
 References

1. JOHNSON, S. M., “Optimal Two and Three Stage Production Schedules with Set-up time
Included”. Nav. Res. Log. Quart. , Vol. 1, pp. 61-68, (1954).
2. CAMPBELL, H. G. “ A heuristic algorithm for n-job and m-machines sequencing
problem”. Management Science, Vol. 16, pp. 630-673, (1970).
3. BAKER, KENNETH R., “ Sequencing and Scheduling”. John Wiley & Sons Inc.(1998),
ISBN: 0-9639746-1-0.
4. GUPTA, J.N.D. , “A functional heuristics algorithm for the flow-shop scheduling
problem”. Operation Research, Vol. 22, No. 1, pp.39- 48, (1971).
5. FRENCH, S. “Sequencing and Scheduling”. An introduction to the mathematics of the
job shop, Ellis Horwood series, (1982).
6. DANNENBRING, D. G. , “ An evaluation of flowshop sequencing heuristics”. Management
Science, Vol. 23, No. 11, pp. 1174-1182, (1977).
7. PINEDO, MICHEAL and BHASKARAN, K. “ Dispatching”. Handbook of Industrial
Engineering, pp. 2182-2198, (1992).
8. BERA, H., “To minimize waiting time of jobs, idle time of machines and total elapsed time
for N jobs on M machines”. 2nd Joint international conference on Production Engineering,
Leicester Polytechnis, pp. 290-299, (1983).
9. EZAT, A. M. and EL BARADIE, M. “ Flow-Shop Scheduling: Effect of Various Priority
Rules on Minimizing Multiple Criteria”. 30th International MATADOR Conference
(UMIST), (1993).
10. TSENG, FAN and STAFFORD, E. ,“Two MILP models for N x M SDST flowshop
sequencing problem ”. Int. J. Prod. Res., Vol. 39, No. 8, 1777 – 1809, (2001).
11. GRAVES, STEPHEN ,“ A Review of Production Scheduling”. Operations Research, Vol.
29, pp. 646-675, (1981).
12. RAJENDRAN, C. and HOLTHAUS, OLIVER, “ A Comparative study of dispatching rules
in dynamic flowshops and jobshops”. European Journal of Operational Research, Vol.
116, pp.156-170, (1999).
13. TAILLARD, ERIC D., “ Some efficient heuristic methods for the flowshop sequencing
problem”. European Journal of Operational Research, Vol.47, pp. 65-74, (1990).
14. PINEDO, MICHEAL, “ Operations scheduling with applications in manufacturing and
services”. Irwin/McGraw-Hill, (1998). ISBN: 0072897791.
15. STAFFORD, E. F., “On the development of a mixed-integer linear programming model for
the standard flowshop”. Journal of the Operational Research Society, Vol. 39, pp. 1163-
1174, (1988).
16. BLAZEWICZ, J., ECKER, KLAUS H., SCHMIDT, GUENTER and WEGLARZ, J.,
“Scheduling in Computer and Manufacturing Systems”. Springer Verlag, Berlin (1994).
ISBN: 3540580492.
17. MEZIANE, VADERA, KOBBACY and PROUDLOVE, “Intelligent Systems in
Manufacturing: Current Developments and Future Prospects”. Integrated Manufacturing
System, Vol. 11, No. 4, pp. 218-238, (2000).
18. CONWAY, R. W., MAXWELL, W.L. and MILLER, I. W. ,“ Theory of Scheduling ”.
Addison Wesley, Reading, Massachusetts, (1967).

View publication stats