V. Modrák, R. S.

Pandian Algoritam planiranja operacija "flow shop" u cilju smanjivanja vremena izvršenja kod problema n-poslova i m-strojeva

ISSN 1330-3651
UDC/UDK 658.514:004.421.2

FLOW SHOP SCHEDULING ALGORITHM TO MINIMIZE COMPLETION TIME

FOR n-JOBS m-MACHINES PROBLEM
Vladimír Modrák, R. Sudhakara Pandian
Original scientific paper
In multi stage job problems, simple priority dispatching rules such as shortest processing time (SPT) and earliest due date (EDD) can be used to obtain solutions
of minimum total processing time, but may not sometimes give sequences as expected that are close to optimal. The Johnson's algorithm is especially popular
among analytical approaches that are used for solving n-jobs, 2-machines sequence problem. In this paper the presented algorithm is based on converting an m-
machine problem to a 2-machine problem. Based on testing and comparison with other relevant methods, the proposed algorithm is offered as a competitive
alternative for practical application when solving n-jobs and m-machines problems.

Keywords: CDS heuristics, flow shop, genetic algorithm, make-span, slope index

Algoritam planiranja operacija "flow shop" u cilju smanjivanja vremena izvršenja kod problema n-poslova i m-strojeva

Izvorni znanstveni članak

U problemima posla s više faza, mogu se koristiti jednostavna prioritetna dispečerska pravila kao što su najkraće vrijeme obrade (PT) i najraniji datum
dospijeća (EDD) za dobivanje rješenja najmanjega ukupnog vremena obrade. Međutim, ona ponekad ne daju slijed za koji se očekuje da je blizu optimalnom.
Johnsonov algoritam je posebno popularan među analitičkim pristupima koji se koriste za rješavanje problema slijeda n-poslova i 2-stroja. Algoritam prikazan
u ovom radu se temelji na pretvaranju problema m-strojeva u problem 2-stroja. Na temelju ispitivanja i usporedbe s drugim relevantnim metodama, predloženi
algoritam se nudi kao konkurentna alternativa za praktičnu primjenu pri rješavanju problema n-poslova i m-strojeva.

Ključne riječi: CDS heuristika, flow shop, genetski algoritam, indikator prioriteta, vrijeme izvršenja posla

1 reduce the number of possible schedules it is reasonable to

Introduction assume that all machines process jobs in the same order [5].
Uvod The deterministic job shop scheduling problem consists of a
finite set J of n jobs to be processed on a finite set M of m
In a shop floor of the industry, the routings which are machines. Each job, Ji, must be processed on every machine
based upon the jobs that need to be processed on different in its routing consisting of mi operations Oi1, Oi2..., Oim
machines are one among the major activities and therefore performed in order.
the resource requirements are not based upon the quantity as The proposed algorithm for minimizing completion
in a flow shop but rather the routings for the products being time is determined for a classical static and deterministic
produced. However, both job shop and flow shop permutation flow shop scheduling problem (PFSSP) with n
production cope with a scheduling problem to find a jobs and m machines which is viewed as sequence problem.
feasible sequence of jobs on given machines with the In the classical flow-shop sequencing and scheduling
objective of minimising some function of the job problem, queues of jobs are allowed at any of m machines in
completion times. Job completion time (make-span) can be processing sequence based on assumption that jobs may
defined as the time span from material availability at the wait on or between the machines [6]. In this study, the
first processing operation to the completion at the last objective function for the PFSS problem corresponds to the
operation [1]. Johnson [2] has shown that, in a 2-machines minimization of the make-span when idle time is allowed on
flow shop, an optimal sequence can be constructed. It was machines.
demonstrated later that m-machine flow shop scheduling
problem (FSSP) is strongly NP-hard for m≥ 3 [3]. FSSPs 2
can be divided into two main categories: dynamic and static. Literature review
The dynamic flow shop considered is one where jobs arrive Pregled literature
continuously over time. The static flow shop-sequencing
and scheduling problem denotes the problem of The flow-shop problem with make-span (cmax) criterion
determining the best sequence of jobs on each machine in can be denoted as either n/m/F/cmax or F//cmax, where both
the flow shop. The criterion of optimality in a flow shop are related to an n-jobs and m-machines problem. This
sequencing problem is usually specified as minimization of notation was firstly suggested by Conway et al. [7] and until
make-span that is defined as the total time to ensure that all now is handy. Pinedo [8] introduced the term Permutation
jobs are completed on all machines. If there are no release Flow-shop Problem (PFSP) in which the processing
times for the jobs then the total completion time equals the sequence on the first machine is maintained throughout the
total flow time. In some cases for calculating the completion remaining machines. Accordingly, the make-span criterion
times specific constraints are assumed. For example, such a is denoted as F/prmu/Cmax. Solution methods for flow shop
situation in the FSSP arises when no idle time is allowed at scheduling range from heuristics developed by Palmer [9],
machines. This constraint creates an important practical Campbell et al. [10], and Dannenbring [11] to more
situation that arises when expensive machinery is employed complex techniques such as branch and bound [12], tabu
[4]. The general scheduling problem for a classical shop search [13, 14, 15], genetic algorithms [16, 17] shifting
m
flow gives rise to (n!) possible schedules. With the aim to bottleneck procedure [18], and ant colony algorithm [19].

Technical Gazette 17, 3(2010), 273-278 273

Flow shop scheduling algorithm to minimize completion time for n-jobs m-machines problem V. Modrák, R. S. Pandian

The concept of a slope index in prioritizing jobs was which is used to find out minimum make-span while 2-
first introduced by Page [20]. Later on Palmer [9] adopted machine production schedules are included. The step by
this idea and proposed the slope index to be utilized for job step algorithm is given in section 3.1
sequencing in the m-machine flow shop problems. A simple
heuristic extension of Johnson's rule to m-machines flow 3.1
shop problem was proposed by Campbell et al [10]. This The algorithm description
extension is known in the literature as the Campbell, Dudek, Opis algoritma
and Smith (CDS) heuristic. Its principle relies on
constructing at most (m-1) different sequences from which Step1. Find out the sum of processing time of n jobs in
the best sequence is chosen. Each sequence corresponds to machine M1.
the application of Johnson's rule on a new 2-machines
Repeat Step 1 for machines j=1, 2, 3, …, m.
problem. CDS heuristics will also be used in this study to
Step 2. Make two groups from m machines in such a
compare solutions of the same PFSPs with the proposed
way that
algorithm.
Another approach to obtain minimum idle time based x m
on the optimization of idle time at the last machine is
presented in [21]. Nawaz et al. [22] proposed that a job with å Ti å Ti ® minimum . (1)
j =1 j = x +1
longer total processing time should have higher priority in
the sequence. More complex heuristics was applied by
Step 3. Find out the total number of machines in each
Ogbu et al. [23] by using simulated annealing and by
group.
Taillard [24] by applying tabu search algorithm for make-
Let the number of machines in Group I = a, and the
span minimization. Nagar et al. [25] proposed a combined
number of machines in Group II = b.
branch-and-bound and genetic algorithm based procedure
Step 4. Calculate total operational time T of jobs in each
for a flow shop scheduling problem with objectives of mean
group using the formula:
flow time and make-span minimization. Similarly, Neppalli
a) for the Group I and Job (J1)
et al. [26] used genetic algorithms in their approach to solve
the 2-machine flow shop problem with the objective of TJI1 = (a × t11 ) + (a - 1) × t12 + (a - 2) × t13 + ..... + (1 × t1a ).
minimizing make-span and total flow time. An atypical
Similarly calculate these values for jobs J2, J3, Jn.
method based on an artificial immune system approach that
was inspired by vertebrate immune system was presented b) for the Group II and Job (J1)
by Engin and Doyen [27]. They used the proposed method TJII1 = (b × t1m ) + (b - 1) × t1m -1 + (b - 2) × t1m - 2 + ..... + (1 × t1a +1 ).
for solving the hybrid flow shop scheduling problem with
minimizing maximum completion times. Similarly calculate these values for jobs J2, J3, Jn.
Even though the various studies suggested many
approaches, it is difficult to find the simplest approach to Step 5. Tabulate these values in two rows.
find an optimal sequence for solving the n-jobs and m- Step 6. Apply final step of Johnson's rule to find out the
machines flow shop scheduling problem. best sequence.
In the future, scheduling approaches in this Step 7. Calculate the make-span time for the sequence
manufacturing area will need to take also market obtained in step 6.
developments into consideration, especially the new Step 8. Store the results.
manufacturing technology and advanced production
control systems that will constrain the overall structure of 3.2
the flow-shop manufacturing operations [28]. Keeping this The algorithm illustration
in mind, scheduling algorithms to minimize make-span for Ilustracija algoritma
n-Jobs m-Machines Problem with simplest steps will be
always needful. To evaluate the proposed algorithm the following 6-
jobs and 5-machines problem from the real life has been
3 used. Input values for the calculation of total operational
The proposed approach to multi stage flow shop time T of jobs in each group are shown in Tab. 1.
sequencing
Predloženi pristup za slijed operacija u višefaznom "flow Table 1 Illustration for the problem of size 6 machines × 5 jobs
Tablica 1. Ilustracija za problem veličine 6 strojeva × 5 poslova
shop"
j\i J1 J2 J3 J4 J5
M1 1 1,5 1,5 1 1
In the multi stage sequencing problem, the following
M2 0,5 0,75 0,75 0,5 0,5
assumptions are made. M3 0,5 1 0,5 0,5 0,5
! There are n number of jobs (J) and m number of M4 0,5 1 0,5 0,5 0,5
machines (M). M5 0,1 0,5 0,2 0,1 0,1
! The order of sequence of operations in all machines is M6 0,2 0,3 0,3 0,1 0,1
the same.
! The setup time is not considered for calculating make-
span time. In the above table, each row represents machine j and
each column represents job i. The processing time of an
operation of the jobs is mentioned in each cell and denoted
The proposed approach works with simple steps as as tij.
given in section 3.1. The optimum sequence is found out in The sum of the processing time of all 5 jobs in each
step 7 that adopts the method of Johnson's algorithm [2], machine is calculated in column Ti as shown in Tab. 2 and

274 Tehnički vjesnik 17, 3(2010), 273-278

V. Modrák, R. S. Pandian Algoritam planiranja operacija "flow shop" u cilju smanjivanja vremena izvršenja kod problema n-poslova i m-strojeva

Figure 1 Gant chart for the criteria of minimum make-span and minimum process interruptions
Slika 1. Gantov dijagram za kriterije minimalnog vremena izvršenja i minimum prekida procesa

Tab. 3. Two groups are formed based on the formula as Table 4 The sum of values of two groups
Tablica 4. Zbroj vrijednosti dviju skupina
given below.
Groups\jobs J1 J2 J3 J4 J5
a m
TIJ 2,5 3,75 3,75 2,5 2,5
TIIJ
å Ti å Ti ® minimum (2)
2,6 5,7 3,3 2,2 2,2
j =1 j = a +1
(a = the arbitrary value from 1 to 5) As per the step 6 of the algorithm, the best sequences
obtained in this method are J1-J2-J3-J5-J4 (or) J1-J2-J3-J4-
2 6 J5.
å Ti - å Ti = 9 - 8 (3) The make-span, when idle time is allowed on machines,
j =1 j =3
is calculated for the J1-J2-J3-J5-J4 sequence (see Tab.5)
since both sequences in the given case bring identical
Thus, the total number of machines in each group is scheduling results.
identified.
The number of machines in Group I (Tab. 2), a = 2 (M1 Table 5 Proposed method J1-J2-J3-J5-J4
and M2 are in Group-I, noted as I). Tablica 5. Predložena metoda J1-J2-J3-J5-J4
The number of machines in Group II (Tab. 3), b = 4 j M1 M2 M3 M4 M5 M6
(M3, M4, M5 and M6 are in Group-II, noted as II). i In Out In Out In Out In Out In Out In Out
J1 0 1 1 1,5 1,5 2 2 2,5 2,5 2,6 2,6 2,8
J2 1 2,5 2,5 3,25 3,25 4,25 4,25 5,25 5,25 5,75 5,75 6,05
Table 2 Group I consisting of two machines J3 2,5 4 4 4,75 4,75 5,25 5,25 5,75 5,75 5,95 6,05 6,35
Tablica 2. I. skupina koja se sastoji od dva stroja
J5 4 5 5 5,5 5,5 6 6 6,5 6,5 6,6 6,6 6,7
j\i J1 J2 J3 J4 J5 Ti Σ Ti J4 5 6 6 6,5 6,5 7 7 7,5 7,5 7,6 7,6 7,7
M1 1 1,5 1,5 1 1 6
9
M2 0,5 0,75 0,75 0,5 0,5 3
With the aim to combine the criterion for calculating the
minimum make-span schedules when idle time is allowed
Table 3 Group II consisting of four machines
Tablica 3. II. skupina koja se sastoji od četiri stroja
on machines along with the criterion for minimum process
j\i J1 J2 J3 J4 J5 Ti Σ Ti
interruptions it is possible to create job schedules by the
M3 0,5 1 0,5 0,5 0,5 3
manner shown in Gant chart in Fig. 1.
M4 0,5 1 0,5 0,5 0,5 3
8 4
M5 0,1 0,5 0,2 0,1 0,1 1
M6 0,2 0,3 0,3 0,1 0,1 1 Comparison with Benchmark Algorithms
Usporedba s repernim algoritmima
Subsequently, for the identified groups I and II the
I II To compare the proposed algorithm with the
values of T Ji and T Ji (for i=1 to n) are calculated for all five
benchmark algorithms, the next three distinct algorithms
jobs. are used: CDS heuristics, Slope index method and Genetic
TIJ1 = (2×1,0) + 0,50 = 2,5 Algorithm. The make-spans for CDS method and Slope
TIJ2 = (2×1,5) + 0,75 = 3,75 algorithm are also calculated and displayed in Tab. 6 and
TIJ3 = (2×1,5) + 0,75 = 3,75 Tab. 7. The sequence obtained by using GA for the same
TIJ4 = (2×1,0) + 0,50 = 2,5 PFSS problem equals the sequence calculated by the
TIJ5 = (2×1,0) + 0,50 = 2,5 proposed method. Moreover these four methods have been
TIIJ1 = (4×0,2)+ (3×0,1) + (2×0,5) + 0,5 = 2,6 employed for finding the best sequence with the other four
problems to achieve more reliable results. For this purpose
TIIJ2 = (4×0,3) + (3×0,5) + (2×1,0) + 1,0 = 5,7
we selected flow shop problems, which are shown in Fig. 2
TIIJ3 = (4×0,3) + (3×0,2) + (2×0,5) + 0,5 = 3,3 (a-d). The results obtained with the benchmark methods are
TIIJ4 = (4×0,1) + (3×0,1) + (2×0,5) + 0,5 = 2,2 compared and shown in Tab. 8.
TIIJ5 = (4×0,1) + (3×0,1) + (2×0,5) + 0,5 = 2,2
I II
The T Ji and T Ji values are tabulated as shown in Tab. 4.

Technical Gazette 17, 3(2010), 273-278 275

Flow shop scheduling algorithm to minimize completion time for n-jobs m-machines problem V. Modrák, R. S. Pandian

4.1 4.2
CDS heuristics Slope index method
CDS heuristika Metoda indikatora prioriteta

As outlined above, the CDS heuristics algorithm [9] is A heuristic has been developed by Palmer [9] in an
basically an extension of the Johnson's algorithm. The focus effort to use Johnson's rule for m≥ 3, since for m=2, this
of the heuristic is the minimization of make-span in a algorithm is slightly different from Johnson's algorithm.
deterministic flow shop problem. The CDS heuristic forms The idea of this procedure is to give priority to some jobs so
in a simple manner a set of an m-1 artificial 2-machine sub- that the jobs with the processing times that tend to increase
problem for the original m-machine problem by summing from machine to machine will receive higher priority, while
the processing times in a manner that combines M1, M2,..., the jobs with the processing times that tend to decrease from
Mm-1 to pseudo machine 1 and M2, M3,... Mm to pseudo machine to machine will receive lower priority.
machine 2. Finally, each of the 2-machine sub-problems is The slope index (SI) for job i is calculated as:
then solved using the Johnson's 2-machines algorithm. The
best of the sequence is selected as the solution to the original m
m-machine problem. SI i = å (2 j - m - 1)tij ,i = 1, 2 ,.....n. (4)
For the given flow shop problem of size 6×5 as given in j =1

Tab. 1, using this heuristic the J2-J3-J1-J5-J4 sequence has

been calculated and the make-span calculation is displayed Then a permutation sequence is determined by ordering
in Tab. 6. the jobs in no increasing order of SIi such as:

j\i J1 J2 J3 J4 SI i1 ³ SI i 2 ³ ....... ³ SI in . (5)

M1 24 61 22 21
M2 7 9 8 6 For the original flow shop 5-jobs and 6-machines
M3 7 5 6 8 problem as given in Tab. 1, using this heuristic the J1-J4-J5-
M4 29 15 14 32 J2-J3 sequence has been calculated and the make-span
(a) calculation is displayed in Tab. 7.
j\i J1 J2 J3 J4 J5 J6 J7
M1 3 2 4 5 1 3 5 Table 7 Slope index method J1-J4-J5-J2-J3
M2 5 5 8 7 2 5 2 Tablica 7. Metoda indikatora prioriteta J1-J4-J5-J2-J3
M3 7 8 1 6 8 4 8 j M1 M2 M3 M4 M5 M6
M4 1 1 6 1 4 6 4 i In Out In Out In Out In Out In Out In Out
M5 6 6 7 8 6 8 6 J1 0 1 1 1,5 1,5 2 2 2,5 2,5 2,6 2,6 2,8
M6 9 7 9 4 7 1 3 J4 1 1,5 1,5 2 2 2,5 2,5 3 3 3,1 3,1 3,2
J5 1,5 2,5 2,5 3 3 3,5 3,5 4 4 4,1 4,1 4,2
M7 4 9 1 3 4 2 2
J2 2,5 4 4 4,8 4,8 5,8 5,8 6,8 6,8 7,3 7,3 7,6
(b) J3 4 5,5 5,5 6,3 6,3 6,8 6,8 7,3 7,3 7,5 7,6 7,9
j\i J1 J2 J3 J4
M1 7 6 5 8
M2 5 6 4 3 4.3
M3 2 4 5 3 Genetic Algorithm (GA)
M4 3 5 6 2 Genetski algoritam (GA)
M5 9 10 8 6
(c) The Genetic Algorithm is a probabilistic approach
j\i J1 J2 J3 J4 J5 J6 J7 which deals with probability [16]. In this paper GA is used
M1 5 2 4 2 5 2 1 to search for optimal solution and the results obtained from
M2 1 1 5 1 4 8 5 the GA are compared with the results of the proposed
M3 4 2 2 5 8 5 4 algorithm as shown in Tab. 8. From this table it is inferred
M4 5 5 3 4 9 5 2 that the proposed approach gives as good results as those of
M5 8 4 5 6 2 6 4 the well known meta-heuristics, GA [17]. This shows the
M6 2 5 8 3 4 5 5 ability of the proposed algorithm with simpler steps. The
M7 4 3 2 2 8 8 2 following steps shown in Fig. 3 provide the pseudo code of
M8 8 2 5 2 5 5 8 GAused for this purpose.
(d)
Figure 2 Input data set for testing flow shop problems
(operational time in hours)
Slika 2. Niz ulaznih podataka za ispitivanje problema protočne radionice
(operativno vrijeme u satima)

Table 6 Make-span calculation for the J2-J3-J1-J5-J4 sequence

Tablica 6. Izračun vremena izvršenja za slijed J2-J3-J1-J5-J4
j M1 M2 M3 M4 M5 M6
i In Out In Out In Out In Out In Out In Out
J2 0 1,5 1,5 2,25 2,25 3,25 3,25 4,25 4,25 4,75 4,75 5,05 Figure 3 The pseudo code of GA
Slika 3. Pseudo kod genetskog algoritma
J3 1,5 3 3 3,75 3,75 4,25 4,25 4,75 4,75 4,77 5,05 5,35
J1 3 4 4 4,5 4,5 5 5 5,5 5,5 5,6 5,6 5,8
J5 4 5 5 5,5 5,5 6 6 6,5 6,5 6,6 6,6 6,7
J4 5 6 6 6,5 6,5 7 7 7,5 7,5 7,6 7,6 7,7

276 Tehnički vjesnik 17, 3(2010), 273-278

V. Modrák, R. S. Pandian Algoritam planiranja operacija "flow shop" u cilju smanjivanja vremena izvršenja kod problema n-poslova i m-strojeva

5 27(1999), 219–239.
[7] Conway, R. W.; Maxwell, W. L.; Miller, L. W. Theory of
Discussion and conclusion Scheduling,Addison-Wesley: Reading, MA, 1967.
Rasprava i zaključak [8] Pinedo, M. Scheduling: Theory, Algorithms and Systems.
Prentice Hall, New Jersey, second edition, 2002.
The present study deals with sequence-dependent [9] Palmer, D. S. Sequencing jobs through a multi-stage process
operations, the sequencing problem which is quite common in the minimum total time – a quick method of obtaining a
in many industries. The main idea is to minimize the make- near optimum, Operations Research. Q. 16(1965), 101-107.
span time thus reducing the idle time of both jobs and [10] Campbell, H. G.; Dudek, R. A.; Smith, M. L. A Heuristic
machines since these criteria are often applied for Algorithm for the n-Job, m-Machine Sequencing Problem,
Management Science, 16, 10(1970), 630-637.
operational decision-making in scheduling. Based on the [11] Dannenbring, David G. An Evaluation of Flow Shop
tested problems it can be concluded that the proposed Sequencing Heuristics, Management Science, 23, 11(1977),
approach produces results comparable with the benchmark 1174-1182.
algorithms as shown in Tab. 8. [12] Brucker, P.; Jurisch, B.; Sievers, B. A branch and bound
algorithm for the job shop scheduling problem. Discrete
Table 8 Comparative results of make-span Applied Mathematics, 49, 1(1994), 109–127.
Tablica 8. Usporedni rezultati vremena izvršenja [13] Gendreau, M.; Laporte, G.; Semet, F. A tabu search heuristic
Number Number Make-span (in hours) for the undirected selective travelling salesman problem,
S. European Journal of Operational Research, Elsevier, 106, 2-
of of CDS Slope Proposed
No GA 3(1998), 539-545.
Machines Parts Algorithm Index Approach
[14] Nowicki, E.; Smutnicki, C. A fast taboo search algorithm for
1 4 4 156,0 157,0 156,0 156,0
the job shop problem. Management Science, 42 ,6(1996),
2 5 4 51,0 51,0 51,0 51,0
797–813.
3 6 5 7,7 7,9 7,7 7,7 [15] Logendran, R.; de Szoeke, P.; Barnard, F. Sequence-
4 7 7 6,7 7,5 6,7 6,7 dependent group scheduling problems in flexible flow shops.
5 8 7 7,1 6,9 6,7 6,7 International Journal of Production Economics, 102 (2006),
66–86.
Many heuristics and meta-heuristics can find quick, [16] Manikas. A.; Chang, Y. L. Multi-criteria sequence-dependent
job shop scheduling using genetic algorithms Computers &
feasible solutions to such sequencing problems that involve
Industrial Engineering, 56 (2009), 179–185.
multiple jobs and machines and sequence-dependent [17] Murata, T.; Ishibuchi. H.; Tanaka, H. Genetic Algorithms for
operations. But, as far as simplicity of the algorithm and Flow shop Scheduling Problems, Computers & Industrial
promising results are concerned, the proposed method is Engineering, 30, 4 (1996), pp. 1061-1071.
more effective than the existing methods. In realistic [18] Balas, E. and A. Vazacopoulos. Guided Local Search with
situation, the proposed algorithm can be used such as it is Shifting Bottleneck for Job Shop Scheduling. Management
without any modification and come out with acceptable Science, 44, 2(1998), 262-275.
results. In that manner the approach can be recommended [19] Blum, C.; Sampels, M. An Ant Colony Optimization
for industries that deal with variety of parts and machines Algorithm for Shop Scheduling Problems. Journal of
with more operations. Mathematical Modelling and Algorithms, 3, 3(2004), 285-
308.
[20] Page, E. S. An Approach to Scheduling of Jobs on the
Acknowledgement Machines, J. Royal Stat. Soc., V 23, (1961), pp. 484-492.
Zahvala [21] Gupta, J. N. D.: Heuristic algorithms for multistage flow shop
This research has been partially supported by the financing scheduling problem,AIIE Transactions, 4, (1)(1972), 11-18.
from VEGAgrant scheme No. 1/4153/07. [22] Nawaz. M.; Enscore, E.; Ham, I. A heuristic algorithm for the
m machine, n job flow shop sequence problem, OMEGA, 11,
1(1983), 91-95.
6 [23] Ogbu, F. A.; Smith, D. K. The application of the simulated
References annealing algorithm to the solution of the n/m/Cmax flow
Literatura shop problem, Computers & Operations Research, 17,
3(1990), 243-253.
[1] Modrák, V.; Modrák, J. Relationships between Batch Sizes, [24] Taillard, E. Some efficient heuristic methods for the flow shop
Sequencing and Lead-times. Proceedings of the 6th sequencing problem, European Journal of Operational
International Conference on Informatics in Control, Research, 47, 1(1990), 65-74.
Automation and Robotics, Intelligent Control Systems and [25] Nagar, A.; Heragu, S. S.; Haddock, J. A combined branch-
Optimization, Milan, Italy, July 2-5, 2009. INSTICC Press and-bound and genetic algorithm based approach for a flow
2009, 380-383. shop-scheduling problem. Annal. Oper. Res., 63(1996),
[2] Johnson, S. M. Optimal Two and Three Stage Production 397–414.
Schedules with Set-Up Times, Included, Naval Research [26] Neppalli, V. R.; Chen, C. L.; Gupta, J. N. D. Genetic
Logistics Quarterly, 1(1954), 61-68. algorithms for the two-stage bicriteria flow shop problem.
[3] Garey, M. R. D.; Johnson, D. S.; Sethi, R. The complexity of Eur. J. Oper. Res., 95(1996), 356–373.
flow shop and job shop scheduling. Mathematics of [27] Engin, O.; Doyen, A. A new approach to solve hybrid flow
Operations Research, 1(1976), 117–129. shop scheduling problems by artificial immune system.
[4] Ruiz, R. Vallada, E.; Fernandez-Mart i nez , C. In. Future Generation Computer Systems, 20(2004), 1083–1095.
Computational Intelligence in Flow Shop and Job Shop [28] Valíček, J.; Hloch, S.; Kozak, D. Surface geometric
Scheduling, Volume 230/2009, Springer Berlin / Heidelberg. parameters proposal for the advanced control of abrasive
(2009) waterjet technology. The International Journal of Advanced
[5] Gupta, J. N. D. Analysis of combinatorial approach to flow Manufacturing Technology. 41, 3-4 (2009), 323-328.
shop scheduling problems. Journal of the Operational
Research Society, 26(1975), 431-440.
[6] Allahverdi, A.; Gupta, J. N. D.; Aldowaisan, T. A review of
scheduling research involving set-up considerations. Omega,

Technical Gazette 17, 3(2010), 273-278 277

Flow shop scheduling algorithm to minimize completion time for n-jobs m-machines problem V. Modrák, R. S. Pandian

Authors' addresses
Adrese autora

Vladimír Modrák
Faculty of Manufacturing Technologies
Technical University of Košice
Bayerova 1, Prešov, Slovakia
[email protected]

R. Sudhakara Pandian
Department of Mechanical Engineering
Kalasalingam University
Krishnankoil-626190, Virudhunagar, India
[email protected]

278 Tehnički vjesnik 17, 3(2010), 273-278