You are on page 1of 6

Advanced Materials Research Vols.

488-489 (2012) pp 1114-1118


Online available since 2012/Mar/15 at www.scientific.net
© (2012) Trans Tech Publications, Switzerland
doi:10.4028/www.scientific.net/AMR.488-489.1114

Optimization Techniques for No-wait Manufacturing Scheduling: A


Review
Sagar U. Sapkal1,a, Dipak Laha1 and Dhiren Kumar Behera1
1
Mechanical Engineering Department, Jadavpur University, Kolkata 700032, India
a
Corresponding author, email: sagar_us@indiatimes.com

Keywords: Manufacturing, Optimization, No-wait scheduling, Heuristics, Metaheuristics

Abstract. This paper deals with a general continuous or no-wait manufacturing scheduling
problem. Due to its applications in advanced manufacturing systems, no-wait scheduling has gained
much attention in both practical and academic fields. Due to its NP-hard nature, most of the
contributions focus on development of approximation based optimization methods or heuristics for
the problem. Several heuristic procedures have been developed to solve this problem. This paper
presents a survey of various methodologies developed to solve no-wait flow shop scheduling
problem with the objective of minimizing single performance measure.

Introduction
Manufacturing scheduling is concerned with setting the timetable for the processing of given set of
jobs on a set of machines in order to optimize a given measure of performance. Manufacturing
scheduling problems may be characterized by different flow patterns of the jobs. The flow pattern
may be same for all the jobs (flow shop), or each job may have its own individual flow pattern (job
shop), or no specified flow pattern may exist (open shop). Recent developments in scheduling
theory have focused on extending the models to include more practical constraints. A brief
classification of manufacturing scheduling is shown in figure 1.
Manufacturing
scheduling

Flow shop Job shop Open shop

Permutation No-wait flow Blocking flow No-idle flow SDST flow Hybrid flow
flow shop shop shop shop shop shop

Fig. 1 Classification of Manufacturing Scheduling

To limit the survey, flow shop scheduling with no-wait constraint is reviewed in this paper. A flow
shop scheduling problem in which each job must be processed until completion without any
interruption either on or between machines is called continuous or no-wait flow shop scheduling
problem (NWFSSP). In a general no-wait flow shop, each of n jobs is processed through m
machines in the same technological order. Each job is to be processed without preemption and
interruption on or between m machines. Applications of NWFSSP can be found in many industries
such as steel industry, plastic moulding industry, process industries, chemical and pharmaceutical
industries, concrete ware production, electronic industry, and food processing industry. Additional
applications can be found in advanced manufacturing environments, such as just-in-time, flexible
manufacturing systems and robotic cells. A detailed presentation of the applications and research on
this problem is given by Hall and Sriskandarajah [1]. In the early research, Wismer [2] and Reddi
and Ramamoorthy [3] modeled NWFSSP as well-known traveling salesman problem (TSP), and
solved the problem using TSP techniques. Based on the Wismer's [2] idea, Van Deman and Baker
[4] and Gupta [5] addressed this problem. Szwarc [6] also addresed a variation of the weighted idle
time problem similar to that considered by Gupta [5]. Panwalkar and Woollam [7] and Panwalker

All rights reserved. No part of contents of this paper may be reproduced or transmitted in any form or by any means without the written permission of TTP,
www.ttp.net. (ID: 130.207.50.37, Georgia Tech Library, Atlanta, USA-15/11/14,05:07:05)
Advanced Materials Research Vols. 488-489 1115

and Woollam [8] considered a special case of the flow shop with no intermediate storage problem
which they denoted as ordered flow shop problem with no waiting and proved that shortest
processing time ordering of jobs minimizes mean flow times. For special cases of this problem,
Adiri and Pohoryles [9] proved some properties of the optimal schedules for a 2-machine problem
and also proved several theorems for polynomial bounded algorithms for m-machine problems. Van
der Veen and Van Dal [10] showed that the problem is solvable when the objective function is
restricted to semi-ordered processing time matrices. Regarding the complexity aspect, NP-hardness
of this problem has been studied by Sahni and Cho [11]. R¨ock [12] showed the problem with more
than two machines is strongly NP-hard. The computational complexity aspect of no-wait shop
scheduling is discussed by Sriskandarajah and Ladet [13]. Along with the computational
complexity, Goyal and Sriskandarajah [14] have also mentioned approximate algorithms for the
problem. In past decades, most research focused on developing heuristics which are generally
classified as constructive heuristics and metaheuristics. The objective criteria generally considered
by researchers are makespan (MS) and total flow time (TFT). Various constructive heuristics and
metaheuristics developed for general NWFSSP with MS and TFT criteria are discussed in
subsequent sections. Finally, conclusion and future scope in this field are reported.

Constructive heuristics on makespan (MS) criterion


Bonney and Gundry [15] proposed an algorithm called the slope matching algorithm based on the
two slopes of each job corresponding to the regression lines which fit respectively the cumulative
process start and end times for NWFSSP. King and Spachis [16] developed some new heuristic
scheduling rules such as single chain and multiple chain heuristics based on minimization of
different delays. Gangadharan and Rajendran [17] proposed two heuristics based on the principles
of minimizing the adjacent inter-job delays and decreasing trend in processing times. Rajendran
[18] developed a heuristic algorithm based on two proposed heuristic preference relations and job
insertion technique. Framinan and Nagano [19] proposed a heuristic based on analogy between
NWFSSP and TSP. Laha and Chakraborty [20] presented a heuristic in which the seed sequence of
the heuristic is generated using the shift neighbourhood mechanism and the final sequence is
obtained by constructing each job incrementally from their proposed initial sequence using a job
insertion principle.

Constructive heuristics on total flow time (TFT) criterion


Rajendran and Chaudhuri [21] used two heuristic relations as the basis for selecting the initial
sequence of jobs for their two heuristic algorithms. The initial sequence of jobs in both the cases is
further improved by using the job insertion approach for developing the final solution. Bertolissi
[22] heuristic is based on a job pair comparison approach. The seed sequence of jobs is created by
arranging the jobs in decreasing number of marks obtained from job pair comparison and then
further improved using the job insertion algorithm described by Rajendran and Chaudhuri [21].
Aldowaisan and Allahverdi [23] proposed six new heuristics by making use of NEH [24] heuristic,
Rajendran and Zeigler [25] heuristic, and pair-wise neighbourhood to their proposed initial
sequence algorithm. Framinan et al. [26] presented a constructive heuristic which is based on
comparing pairs of jobs in order to select a candidate job incrementally for appending it to the
current partial schedule. Then, the best schedule is obtained among the generated sequences by
placing the candidate job at all possible positions in the current partial schedule using the insertion
and swap neighbourhoods.
1116 Key Engineering Materials II

Table1: Constructive heuristics in no-wait flow shop scheduling


References Criterion Basis for development of Methods compared Complexity
heuristics
Bonney & MS Slope matching algorithm Palmer [27], Gupta [28], O(n2)
Gundry [15] , TSP
King & Spachis MS ATSP Single chain and multi chain heuristics O(n2)
[16]
Gangadharan & MS Job insertion Bonney & Gundry [15], King & Spachis O(n2)
Rajendran [17] [16]
Rajendran [18] MS Job insertion Bonney & Gundry [15], King & Spachis O(n2)
[16]
Framinan & MS Farthest Index development strategies and O(n2)
Nigano [19] Insertion Travelling solution construction procedures
Salesman Procedure
Laha & MS Job insertion Gangadharan & Rajendran [17], O(n2)
Chakraborty [20] Rajendran [18], Osman & Potts [29],
Aldowaisan & Allahverdi [30]
Rajendran & TFT Job insertion Bonney & Gundry [15], King & Spachis O(n2)
Chaudhuri [21] [16]
Bertolissi [22] TFT Job insertion Rajendran & Chaudhuri [21], Bonney & O(n2)
Gundry [15]
Aldowaisan & TFT Insertion method, pair- Rajendran & Chaudhuri [21], Chen et al. O(n3)
Allahverdi [23] wise exchange procedure [31]
Framinan et al. TFT Insertion & exchange Rajendran & Chaudhuri [21], Bertolissi O(n3)
[26] neighbourhood. [22], Aldowaisan & Allahverdi [23],
Fink & Voβ [32]

Metaheuristics on makespan (MS) criterion


Aldowaisan and Allahverdi [30] applied simulated annealing (SA) and genetic algorithm (GA) to
NWFSSP and compared with the best two existing constructive heuristics of Gangadharan and
Rajendran [17] and Rajendran [18]. Grabowski and Pempera [33] proposed two local search
algorithms based on traditional descending search (DS) approach, namely, DS and DS with
multimoves (DS+M) and three local search algorithms based on Tabu Search (TS), namely, TS, TS
with multimoves (TS+M), and TS with multimoves with a tuning parameter (TS+MP). Liu et al.
[34] proposed a hybrid particle swarm optimization (HPSO) by incorporating another local search
based on SA with multiple different neighbourhoods guided by an adaptive meta-Lamarckian
learning strategy for the scheduling. Pan et al. [35] presented a discrete particle swarm optimization
(DPSO) algorithm to solve NWFSSP with both the objectives MS as well as TFT. Pan et al. [36]
developed a hybrid discrete particle swarm optimization (HDPSO) algorithm combining with a
local search algorithm based on the insert neighbourhood and showed that the performance of the
HDPSO algorithm is competitive in terms of the quality of solutions and CPU times. An improved
iterated greedy algorithm (IIGA) is proposed by Pan et al. [37] to solve this problem. Qian et al.
[38] proposed a hybrid differential evolution (HDE) by combining DE with local search and
showed that the performance of the HDE is efficient and effective compared to the best existing
metaheuristics.
Metaheuristics on total flow time (TFT) criterion
Chen et al. [31] developed a GA based heuristic and have shown to perform better than that
obtained by the job insertion method of Rajendran and Chaudhuri heuristic [21]. Fink and Voβ [32]
considered steepest descent (SD), SA, and TS to discuss the effectiveness of implementing them to
the continuous flow shop scheduling problems. Kumar et al. [39] proposed a Psycho-Clonal
algorithm based on the artificial immune system (AIS) approach combined with Maslow’s need
hierarchy theory. Pan et al. [35] also applied their proposed HDPSO with the variable
neighbourhood descent for NWFSSP with the TFT criterion and they reported that their algorithm is
efficient and effective due to the inclusion of several speed-up methods for the swap and insert
neighbourhoods.
Advanced Materials Research Vols. 488-489 1117

Table 2: Metaheuristics in no-wait flow shop scheduling


References Criterion Metaheuristics Methods compared
Aldowaisan & MS SA, GA Gangadharan & Rajendran [17], and Rajendran [18]
Allahverdi [30]
Grabowski & MS DS, DS + M, TS, TS + Schuster & Framinan [40]
Pempera [33] M, TS+MP
Liu et al. [34] MS HPSO Rajendran [18], and Schuster & Framinan [40]
Pan et al. [35] MS DPSO Rajendran [18], Schuster and Framinan [40], and
Grabowski & Pempera [33]
Pan et al. [36] MS HDPSO Liu et al. [34], Pan et al. [35]
Pan et al. [37] MS IIGA Rajendran [18], and Grabowski & Pempera [33]
Qian et al. [38] MS HDE Rajendran [18], Grabowski & Pempera [33], Osman
& Potts [29], and Liu et al. [34]
Chen et al. [31] TFT GA Rajendran & Chaudhuri [21]
Fink & Voß [32] TFT SD, SA, TS Cheapest insertion heuristic
Kumar et al. [39] TFT AIS Rajendran & Chaudhuri [21], Chen et al. [31], and
Aldowaisan and Allahverdi [23]
Pan et al. [35] TFT HDPSO Fink & Voß [32]

Concluding remarks
In this paper we reviewed heuristic methods for solving general NWFSSP with single objective
criterion. These heuristic methods are generally classified as constructive heuristics and
metaheuristics. Constructive heuristics may be attractive from a practical point of view, since these
are rather easy to understand and allow a straightforward implementation. They also frequently
produce near-optimal solutions in reasonable computational efforts and do not require calibration.
Metaheuristics can usually produce better effectiveness than constructive heuristics but these
procedures require much computation time to be acceptable for practice. It is desirable to obtain
both good effectiveness and high efficiency. This paper can be used as a reference of past
contributions in no-wait flow shop scheduling. Future research should examine the application of
constructive heuristics, metaheuristics and also some hybrid heuristics for further improvement of
methods in this field.

References
[1] N.G. Hall and C. Sriskndarajah: Operations Research. Vol. 44-3 (1996), p. 510
[2] D.A. Wismer: Operations Research. Vol. 20 (1972), p. 689
[3] S.S. Reddi and C.V. Ramamoorthy: Operational Research Quarterly. Vol. 23 (1972), p. 323
[4] J.M. Van Deman and K.R. Baker: AIIE Transactions. Vol. 6 (1974), p. 28
[5] J.N.D. Gupta: Naval Research Logistic Quarterly. Vol. 23-2 (1976), p. 235
[6] W. Szwarc: Naval Research Logistic Quarterly. Vol. 28 (1981), p. 665
[7] S.S. Panwalkar and C.R. Woollam: J of Operational Research Soc. Vol. 30 (1979), p. 661
[8] S.S. Panwalker and C.R. Woollam: J of Operational Research Soc. Vol. 31 (1980), p. 1039
[9] I. Adiri and D. Pohoryles: Naval Research Logistic Quarterly. Vol. 29 (1982), p. 495
[10] J.A.A. Van der Veen and R. Van Dal: J of the Operational Research Soc. Vol. 42 (1991), p.
971
[11] S. Sahni and Y. Cho: Math. Opns. Res. Vol. 1 (1979), p. 448
[12] H. R¨ock: J of Associate Computer Machinery. Vol. 31 (1984), p. 336
1118 Key Engineering Materials II

[13] C. Sriskandarajah and P. Ladet: European J of Operational Research. Vol. 24-3 (1986), p. 424
[14] S.K. Goyal and C. Sriskandarajah: Operations Research. Vol. 25-4 (1988), p. 220
[15] M.C. Bonney and S.W. Gundry: Operational Research Quarterly. Vol. 27 (1976), p. 869
[16] J.R. King and A.S. Spachis: Int J of Production Research. Vol. 18 (1980), p. 343
[17] R. Gangadharan and C. Rajendran: Int J of Production Economics, Vol. 32-3 (1993), p. 285
[18] C. Rajendran: J of Operational Research Soc. Vol. 45 (1994), p. 472
[19] J.M. Framinan and M.S. Nagano: J of Materials Processing Tech. Vol. 197 (2008), p. 1
[20] D. Laha and U.K. Chakraborty: Int J of Advanced Mfg Tech. Vol. 41 (2009), p. 97
[21] C. Rajendran and D. Chaudhuri: Naval Research Logistic Quarterly. Vol. 37 (1990), p. 695
[22] E. Bertolissi: J of Material Processing Tech. Vol. 107 (2000), p. 459
[23] T. Aldowiasan and A. Allhverdi: Omega - Int J of Management Science. Vol. 32 (2004), p. 345
[24] M. Nawaz, E.E. Enscore Jr and I. Ham: Omega-Int J of Management Sci. Vol. 11-1 (1983), p.
91
[25] C. Rajendran and H. Ziegler: European J of Operational Research. Vol. 103 (1997), p. 129
[26] J.M. Framinan, M.S. Nagano and J.V. Moccellin: Int J of Adv Mfg Tech. Vol. 46 (2010), p.
1049
[27] D.S. Palmer: Operational Research Quarterly. Vol. 16 (1965), p. 101
[28] J.N.D. Gupta: Operational Research Quarterly. Vol. 22 (1971), p. 39
[29] I.H. Osman, C.N. Potts: Omega-Int J of Management Science. Vol. 17 (1989), p. 551
[30] T. Aldowaisan and A. Allahverdi: Computers and Operations Res. Vol. 30-8 (2003), p. 1219
[31] C. Chen, V. Neppalli and N. Aljaber: Computers and Industrial Engg. Vol. 30 (1996), p. 919
[32] A. Fink and S. Voß: European J of Operational Research, Vol. 151 (2003), p. 400
[33] J. Grabowski and J. Pempera: Computers and Operations Research. Vol. 32-8 (2005), p. 2197
[34] B. Liu, L. Wang and Y.H. Jin: Int J of Advanced Mfg Tech. Vol. 31 (2007), p. 1001
[35] Q. K. Pan, M.F. Tasgetiren and Y.C. Liang: Computers and Opns Res. Vol. 35 (2008), p. 2807
[36] Q.K. Pan, L. Wang, M.F. Tasgetiren and B.H. Zhao: Int J of Adv Mfg Tech. Vol. 38 (2008), p.
337
[37] Q.K. Pan, L. Wang and B.H. Zhao: Int J of Advanced Mfg Tech. Vol. 38-7 (2008), p. 778
[38] B. Qian, L. Wang, R. Hub, D.X. Huang and X. Wang: Comp & Ind Engg, Vol. 57-3 (2009), p.
787
[39] A. Kumar, A. Prakash, R. Shankar and M.K. Tiwari: Expert Sys & Applns. Vol. 31 (2006), p.
504
[40] C.J. Schuster and J.M. Framinan: Operations Research Letters. Vol. 31 (2003), p. 308
Key Engineering Materials II
10.4028/www.scientific.net/AMR.488-489

Optimization Techniques for No-Wait Manufacturing Scheduling: A Review


10.4028/www.scientific.net/AMR.488-489.1114

DOI References
[1] N.G. Hall and C. Sriskndarajah: Operations Research. Vol. 44-3 (1996), p.510.
http://dx.doi.org/10.1287/opre.44.3.510
[2] D.A. Wismer: Operations Research. Vol. 20 (1972), p.689.
http://dx.doi.org/10.1287/opre.20.3.689
[3] S.S. Reddi and C.V. Ramamoorthy: Operational Research Quarterly. Vol. 23 (1972), p.323.
http://dx.doi.org/10.1057/jors.1972.52
[4] J.M. Van Deman and K.R. Baker: AIIE Transactions. Vol. 6 (1974), p.28.
http://dx.doi.org/10.1080/05695557408974929
[5] J.N.D. Gupta: Naval Research Logistic Quarterly. Vol. 23-2 (1976), p.235.
http://dx.doi.org/10.1002/nav.3800230206
[6] W. Szwarc: Naval Research Logistic Quarterly. Vol. 28 (1981), p.665.
http://dx.doi.org/10.1002/nav.3800280413
[9] I. Adiri and D. Pohoryles: Naval Research Logistic Quarterly. Vol. 29 (1982), p.495.
http://dx.doi.org/10.1002/nav.3800290311
[10] J.A.A. Van der Veen and R. Van Dal: J of the Operational Research Soc. Vol. 42 (1991), p.971.
http://dx.doi.org/10.1038/sj/jors/0421105
[11] S. Sahni and Y. Cho: Math. Opns. Res. Vol. 1 (1979), p.448.
http://dx.doi.org/10.1287/moor.4.4.448
[12] H. R¨ock: J of Associate Computer Machinery. Vol. 31 (1984), p.336.
http://dx.doi.org/10.1145/62.65

You might also like