Está en la página 1de 13

Int J Adv Manuf Technol (2009) 43:926938

DOI 10.1007/s00170-008-1770-4

ORIGINAL ARTICLE

Multi-objective production scheduling: a survey


Deming Lei

Received: 21 May 2008 / Accepted: 18 September 2008 / Published online: 3 October 2008
# Springer-Verlag London Limited 2008

Abstract The real life scheduling problems often have


several conflicting objectives. The solutions of these
problems can provide deeper insights to the decision
maker than those of single-objective problems. However, the literature of multi-objective scheduling is notably
sparser than that of single-objective scheduling. Since
the survey paper on multi-objective and bi-objective
scheduling was conducted by Nagar et al. in 1995, there
has been an increasing interest in multi-objective
production scheduling, especially in multi-objective
deterministic problem. The goal of this paper was to
provide an extensive review of the literature on the
scheduling problems with multiple objectives in the past
13 years. This paper also presents some problems
receiving less attention than the others.
Keywords Scheduling problem . Multiple objectives .
Deterministic scheduling . Uncertainty

1 Introduction
Multi-objective scheduling problem (MOSP) is the one
with multiple conflicting objectives which presents some
difficulties related to objectives. If all objectives are
combined into a scalar function by using weights, the

D. Lei (*)
School of Automation, Wuhan University of Technology,
122 Luoshi Road,
Wuhan, Hubei Province, Peoples Republic of China
e-mail: deminglei11@163.com

difficulty is to assign a weight for each objective; if


objectives are optimized concurrently, the problem is to
design the effective search algorithm for some extra steps
and the considerable increment of time complexity.
Before 1995, the conventional techniques are the
main approaches to multi- and bi-objective problems.
These methods can only solve the small size problems.
Meta-heuristics have become the main path to solve
MOSP since 1995. MOSPs have been extensively
investigated, and a number of effective scheduling
algorithms have been occurring; moreover, many published papers aim to provide a set of non-dominated
solutions.
There have been several survey papers over the years.
Nagar et al. [1] provided a good review of the researches
related to regular measures up to early 1990s. They
discussed four types of problems: single machine bicriteria problem, single machine multiple criteria problem,
multiple machine bi-criteria problem, and multiple machine
multiple criteria problem. Tkindt et al. [2] listed more than
100 published papers and focused on three types of
problems: one-machine job shop, parallel-machine job shop,
and flow shop. Hoogeveen [3] paid special attention to
earlinesstardiness scheduling, scheduling with the controllable processing time, simultaneous approximation, and new
models.
In this paper, the scheduling problems are first classified
based on the nature of the problem, shop configuration, and
the description method of uncertainty. The literature on
multi-objective scheduling after 1995 is then extensively
provided. Finally, the characteristics of the previous
researches are summarized and the new trends in scheduling with multiple objectives are discussed.

Int J Adv Manuf Technol (2009) 43:926938

927

2 Definition and classification


2.1 Concepts on Pareto optimality
The general multi-objective optimization problem is of the
form:
max F X f1 X ; f2 X ; :::fK X 
Subject to gi X  0 i 1; 2; :::L

1
n

where X is decision vector, X D R , D is search space,


and gi is a constraint.
The following basic concepts are often used in multiobjective optimization case.
Definition 1: For a decision vector X0 D
1. X0 is said to dominate a decision vector X1 D
(X 0  X 1 ) if and only if




f1 X 0 . . . f1 X 1 i 1; 2;    K; f1 X 0 > f1 X 1 9i 2 f1; 2;    K g:

2. X0 is said to be Pareto optimal if and only if


:9X 1 2 D : X 1  X 0 .
3. Pareto optimal set PS of all Pareto optimal decision
vectors.



PS X 0 2 D :9X 1 2 D; X 1  X 0
2
4. Pareto optimal front PF of all objective function values
corresponding to the decision vectors in PS :


PF F X f1 X ; f2 X ;    fK X jX 2 PF
3
5. X0 is said to be non-dominated regarding a given set if
X0 is not dominated by any decision vectors in the set.
When a decision vector is non-dominated regarding the
whole search space, it is Pareto optimal. Pareto optimal
decision vector cannot be improved in any objectives
without causing degradation in at least one other objective.
A multi-objective optimization algorithm can produce a great
number of non-dominated solutions, and only representatives
of them are stored in an external archive or external
population to save computational resources. The external
archive becomes the inseparable part of many algorithms
such as Pareto archived evolution strategy [4], strength
Pareto evolutionary algorithm (SPEA [5]) and SPEA2 [6].
2.2 Classification
Nagar et al. [1] provided a complete classification scheme
based on four characteristics: nature of the problem, shop

configuration, solution methodology, and performance


measure.
In this paper, the scheduling problems are first classified
into two classes in terms of the nature of problem. The first
class is the deterministic problem in which the processing
constraints and parameters can be ascertained with certainty. The second is the uncertain scheduling problem in which
some processing conditions or parameters cannot be
determined in advance.
Secondly, the deterministic problems are categorized
into four types according to shop configuration. These
types are single machine, parallel machines, flow shop,
and job shop. In parallel-machine shop, a number of oneoperation jobs can be processed on any of machines. In
flow shop, machines are arranged in a serial fashion, and
each job has to pass through each machine. Job shop is a
configuration in which each job has different processing
routes.
Finally, the uncertain scheduling problems are grouped
into two types in terms of the description method of
uncertainty. The first type is fuzzy scheduling problem in
which the processing conditions and parameters are
modeled using fuzzy number. The second is stochastic
scheduling problem in which stochastic variable is used to
indicate the processing constraints and parameters. Figure 1
shows the classification of the problems.

3 Deterministic scheduling: single machine and parallel


machines
As stated in Nagar et al. [1], multi-objective scheduling
researches focus on single machine and parallel machines,
and the main solution approaches are the conventional
techniques such as branch-and-bound and programming
method before 1995. The applications of genetic algorithm
(GA) have notably increased, and the special constraints
have also been considered since 1995. This section
describes multi-objective literature in single-machine and
parallel-machine environment, a summary of which is
shown in Table 1.
3.1 Single machine
Kksalan and Keha [7] considered two problems involving
flow time and the number of tardy jobs, flow time and the
maximum earliness, respectively. They developed a heuristic and a GA approach for the first problem and suggested a
GA for the second.
Gupta and Sivakumar [8] dealt with the problem of
scheduling n independent jobs on a single testing machine
with due dates and sequence-dependent setup times. They
proposed a compromise programming technique to mini-

928

Int J Adv Manuf Technol (2009) 43:926938

Fig. 1 A classification of
scheduling problems

mize the average cycle time and the average tardiness and
to maximize machine utilization.
Azizoglu et al. [9] presented two general procedures
to generate all efficient schedules and select the best
schedule, respectively, for the problem involving the
maximum earliness and the number of tardy jobs. Jolai
et al. [10] proposed a GA for the same problem. Haral
et al. [11] considered bi-criteria scheduling with nontraditional requirements and suggested an experimental
approach and a random key GA to find Pareto optimal
solutions.
Eren and Gner [12] addressed a bi-criteria problem with
sequence-dependent setup times. They first presented an
integer programming model and then developed a special
heuristic and tabu search (TS) for large size problem. Chen
et al. [13] suggested a heuristic and a branch-and-bound
algorithm for the problem with periodic maintenance. The

objective is to minimize the total flow time and the


maximum tardiness.
3.2 Parallel machines
Suresh and Chaudhuri [14] proposed a TS-based approach
to solve the problem involving the maximum completion
time and the maximum tardiness. The parallel machines
considered are of unrelated type, and the jobs require only a
single stage of operation on any one of machines. They
showed that the proposed TS performs better than a
heuristic from literature in terms of quality of solutions
and CPU time. Tkindt et al. [15] studied the resolution of a
bi-criteria scheduling problem and proposed an algorithm to
characterize all strict Pareto optima.
Cochran et al. [16] presented a two-stage multi-population
genetic algorithm (MPGA) for minimizing the makespan, the

Table 1 Researches on multi-objective scheduling on single machine and parallel machines


References

Problem

Objectives

Approach

Special constraints

Kksalan and Keha [7]


Gupta and Sivakumar [8]

Single machine
Single machine

Two
Two

GA
compromise programming technique

Azizoglu et al. [9]


Jolai et al. [10]
Haral et al. [11]
Eren and Gner [12]
Chen [13]
Suresh and Chaudhuri [14]
Tkindt et al. [15]
Cochran et al.[16]
Chang et al. [18]
Chang et al. [21]
Eren [22]

Single machine
Single machine
Single machine
Single machine
Single machine
Parallel machine
Parallel machine
Parallel machine
Parallel machine
Parallel machine
Parallel machine

Two
Two
Two
Two
Two
Two
Two
Two, Three
Two
Two
Two

Heuristic
GA
GA
TS
Heuristic
GA
Heuristic
GA
GA
GA
Heuristic

NO
sequence-dependent setup times
and due dates
NO
NO
Non-traditional
Sequence-dependent setup time
Periodic maintenance
NO
NO
NO
NO
NO
Setup times and remove times

Int J Adv Manuf Technol (2009) 43:926938

total weighted tardiness, and the total weighted completion


time. MPGA is first applied to optimize the first two
objectives and then extended to minimize three objectives.
Computational results show that MPGA outperforms multiobjective genetic algorithm (MOGA [17]).
Chang et al. [18] studied the problem of minimizing
makespan and total tardiness and suggested a two-phase
subpopulation genetic algorithm (SPGA). Computational
results demonstrate that SPGA has better performance than
MOGA [19] and non-dominated sorting genetic algorithm 2
(NSGA2 [20]) in average performance. Chang et al. [21]
modified SPGA using a global archive and an adaptive
strategy for the same problem. They stated that the
modified SPGA tends to outperform NSGA2 and SPEA2,
especially in the large size problems.
Eren [22] considered a bi-criteria m identical parallel
machines problem with a learning effect of setup times and
removal times. They developed a mathematical programming model and three heuristic approaches and showed that
the proposed model is effective in solving problems with up
to 15 jobs and five machines.

4 Deterministic scheduling: flow shop


More than 1,300 papers on various aspects of flow shop
scheduling problem (FSSP) have been published since
1954. More than 50 papers addressed multi-objective flow
shop scheduling problems in the past 13 years, and most of
them are related to permutation flow shop scheduling
problem (PFSSP) and FSSP. Only one paper addressed
hybrid flow shop scheduling problem (HFSSP). This
section describes the multi-objective scheduling literature
in flow shop environments, a summary of which is
provided in Table 2.
As shown in Table 2, the literature on multi-objective
flow shop scheduling has the following features:
1. The conventional techniques, heuristics, and metaheuristics have been applied to solve FSSP with
multiple objectives where meta-heuristics such as GA
have been used in 36 of a total of 53 papers.
2. Only eight papers considered three-objective problem,
and nine papers focus on the problems with the special
constraints such as setup time, learning effect, and
limited buffers, etc. The problems with more than three
objectives and more special constraints have attracted
less attention.
4.1 Permutation flow shop
Danneberg et al. [23] proposed several heuristics for
solving the problem with setup times where the jobs are

929

partitioned into groups or families. Sivrikaya-Serifolu and


Ulusoy [24] compared three brand-and-bound approaches
for bi-criteria two-machine problem.
Ishibuchi et al. [25] modified multi-objective genetic
local search (MOGLS [57]) by choosing only good
individuals as the initial solutions for local search and
assigning an appropriate local search direction to each
initial solution. They demonstrated the importance of
striking a balance between genetic search and local search.
Geiger [26] studied the problem structure of PFSSP with
multiple objectives and investigated the effectiveness of
local search neighborhoods within an evolutionary algorithm (EA) framework.
Rajendran and Ziegler [27] considered the problem with
objectives of the makespan and total flow time. They
presented two ant colony optimization (ACO) algorithms in
which the first extends the ideas of maxmin ant system
(MMAS [28]) and the second is also quite different from
MMAS. Computational results showed that the preposed
algorithms perform better than the existing ACOs such as
MMAS. Pasia et al.[29] investigated the performance of a
Pareto ACO.
Loukil et al. [30] first presented the general context of
multi-objective production scheduling and then developed a
simulated annealing (SA) for single machine, parallel
machines, and permutation flow shop scheduling with
more than one objective. Varadharajan and Rajendran [31]
suggested a two-phase multi-objective simulated annealing
(MOSA), which seeks to obtain non-dominated solutions
through the implementation of a simple probability function
that selects probabilistically a particular objective function.
They presented two variants of the MOSA and showed that
most of the solutions in net non-dominated front are yielded
by these variants.
Armentano and Arroyo [32] presented a TS-based
approach which works with a set of non-dominated
solutions and searches for a better approximation of the
efficient points. They showed that the TS approach is very
competitive when compared with two meta-heuristics from
the literature.
Basseur [33] proposed an adaptive genetic algorithm
(AGA) and discussed several cooperation schemes between
AGA and other methods. Pasupathy et al. [34] developed a
Pareto genetic algorithm with an archive of non-dominated
solutions subjected to a local search for minimizing the
makespan and the total flow time.
Rahimi-Vahed and Mirghorbani [35] presented an
effective particle swarm algorithm (PSO) which exploits
a new concept of the ideal point and a new approach to
specify the superior particles position vector in the
swarm. Guo et al. [36] suggested a discrete particle
swarm optimization in which fitness function considering
both Pareto dominance and the phenotype neighborhood

930

Int J Adv Manuf Technol (2009) 43:926938

Table 2 Researches on multi-objective flow shop scheduling problem


References

Problem

Objectives

Approach

Special constraints

Danneberg et al. [22]


Sivrikaya-Serifolu and Ulusoy [23]
Ishibuchi et al. [24]
Geiger [25]
Rajendran and Ziegler [26]
Pasia et al. [28]
Loukil et al. [29]
Varadharajan and Rajendran [30]
Armentano and Arroyo [31]
Basseur [32]
Pasupathy et al. [33]
Rahimi-Vahed and Mirghorbani[34]
Guo et al. [35]
Qian et al.[36]
Allahverdi [38]
Allahverdi [39]
Allahverdi and Aldowaisan [40]
Allahverdi and Aldowaisan [41]
Allahverdi and Mittenthal [42]
Allahverdi [43]
Allahverdi and Savsar [44]
Allehverdi and Al-Anzi [45]
Yeh and Allahverdi [46]
Rajendran [47]
Ravindran et al. [48]
Chou and Lee [49]
Murata et al. [17]
Nagar et al. [50]
Basseur et al. [51]
Sayin and Karabati [52]
Toktas et al. [53]
Neppalli et al. [54]
Sridhar and Rajendran [55]
Chang et al.[56]
Ishibuchi and Murata [57]
Talbi et al. [58]
Arroyo and Armentano [59]
Brizuela et al. [60]
Ponnambalam et al. [61]
Li and Wang [62]
Eren and Gner [63]
Eren and Gner [64]
Eren and Gner [65]
Tavakkoli-Moghaddam et al. [67]
Tavakkoli-Moghaddam et al. [68]
Li and Zhang [69]
Yagmahan and Yenisey [70]
Tkindt et al. [71]
Ruiz and Allahverdi [72]
Noorul Haq and Rahda Ramanan [73]
Mansouri and Hendizadeh [74]
Liao et al. [75]
Wei et al. [76]

PFSSP
PFSSP
PFSSP
PFSSP
PFSSP
PFSSP
PFSSP
PFSSP
PFSSP
PFSSP
PFSSP
PFSSP
PFSSP
PFSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
FSSP
HFSSP

Two
Two
Two
Up to six
Two
Two
Two, three
Two
Two
Two
Two
Two
Two
Two
Two
Two
Two
Two
Two
Three
Three
Two
Two
Three
Two
Two
Three
Two
Two
Two
Two
Two
Two
Two
Two
Two
Two
Two
Three
Two
Two
Two
Three
Two
Two
Two
Two
Two
Two
Two
Two
Two
Two, four

Heuristics
Branch-and-bound
GA and local search
EA and local search
ACO
ACO and Path Relinking
SA
SA
TS
GA and exact method
GA
PSO
PSO
Differential evolution(DE) and local search
Heuristics
Heuristics
SA, GA
Heuristics, branch-and-bound
Heuristics
Heuristics
Heuristics
SA,ACO,SDE
Heuristic, branch-and-bound
Heuristics
Heuristics
Heuristics
GA
GA and branch-and-bound
Exact method and GA
Branch-and-bound
Branch-and-bound
GA
GA
GA
GA and local search
GA and local search
EA and local search
GA
TSP-GA
Quantum-inspired GA
TS, heuristic
TS, heuristic
TS, heuristic, EDD
HMOIA
MOIA
ACO
ACO
ACO
GA
Neural network
GA, SA
Branch-and-bound
EA

Batch, setup time


NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
Limited buffers
NO
NO
NO
NO
Breakdown
NO
Duplicate station
Assembly machine
NO
NO
NO
Release dates
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
Setup times
Learning effect
NO
NO
NO
NO
NO
NO
NO
NO
Setup times
NO
NO

Int J Adv Manuf Technol (2009) 43:926938

density is used to select the global best position for


particles.
Qian et al. [37] first presented a hybrid differential
evolution (HDE) for the problem with limited buffers
between consecutive machines in which the DE-based
algorithm is applied for exploration in continuous space
and then a local search is used to search in permutationbased solution space. They proved that HDE can converge
to some Pareto solutions with probability 1.
4.2 Flow shop
Allahverdi [38, 39, 43] suggested some heuristics for twomachine or m-machine problems. Allahverdi and Aldowaisan
[40] presented a hybrid SA and a hybrid GA to minimize the
weighted sum of the makespan and the maximum lateness.
Allahverdi and Aldowaisan [41] developed a branch-andbound approach and a heuristic for two- and m-machine
problems, respectively. Allahverdi and Mittenthal [42]
discussed a two-machine problem with random breakdown
and showed that SPT and LPT are optimal when the first or
the second machine suffers random breakdown, respectively.
Allahverdi and Savsar [44] addressed the scheduling
problem in an assembly line with three duplicate stations
and developed two algorithms to the optimal schedules.
Allehverdi and Al-Anzi [45] proposed SA, ACO, and selfadaptive differential evolution for two-stage assembly flow
shop. Yeh and Allahverdi [46] suggested a two-phase hybrid
heuristic and a branch-and-bound approach for three-machine problem.
Rajendran [47] addressed the problem of minimizing the
makespan, the total flow time, and machine idle time. They
first proposed a heuristic for the first two objectives and
then extended the heuristic for all objectives. Ravindran et
al. [48] proposed three heuristic algorithms to minimize the
makespan and the total flow time. Chou and Lee [49]
presented a heuristic for two-machine problem with release
dates of the jobs.
Nagar et al. [50] developed a combined branch-andbound and GA-based algorithm and two modified versions
of it. Basseur et al. [51] presented a two-phase method and
discussed an original cooperation between the exact method
and GA to obtain good solutions of large size problem.
Sayin and Karabati [52] developed a branch-and-bound
procedure to minimize the makespan and the sum of
completion times, which iteratively solves the restricted
single-objective scheduling problems until the set of Pareto
optimal solutions is completely enumerated.
Toktas et al. [53] studied the problem of minimizing the
makespan and the maximum earliness. They developed a
branch-and-bound approach to generate all Pareto optimal
solutions and a heuristic algorithm to produce approximate
optimal solutions. Computational results reveal that the

931

brand-and-bound method can solve problems with up to 25


jobs.
Neppalli et al. [54] suggested two GA-based methods to
minimize the total flow time and the makespan. Sridhar and
Rajendran [55] developed a MOGA for the problem in flow
shop and cellular manufacturing systems with the objectives of minimizing the makespan, the total flow time, and
machine idleness. Murata et al. [17] presented a MOGA to
minimize the objectives of the makespan and the total
tardiness. Chang et al. [56] suggested a SPGA with miming
gene structure.
Ishibuchi and Murata [57] presented an effective
MOGLS called IM-MOGLS in which a local search
procedure is applied to each solution generated by genetic
operations. Talbi et al. [58] provided a simplification of
genetic local search where multi-objective local search is
applied when GA stops. Arroyo and Armentano [59]
designed a MOGLS with features such as preservation of
dispersion in the population, elitism, and use of a parallel
multi-objective local search so as intensify the search in
distinct regions.
Brizuela et al. [60] analyzed how the genetic operators
influence the generation of non-dominated solutions
according to the parental distance. Ponnambalam et al.
[61] suggested a traveling salesman algorithm and GAbased multi-objective algorithm for minimizing the weighted sum of multiple objectives.
Li and Wang [62] proposed a hybrid quantum-inspired
genetic algorithm (HQGA) based on quantum-inspired
genetic algorithm (QGA) and permutation-based genetic
algorithm (PGA). QGA is for exploration in discrete 01
hyperspace by using the updating operator of quantum gate
and genetic operators of Q-bit. PGA is used for both
exploration in permutation-based scheduling space and
exploitation for good scheduling solutions. They stated that
HQGA can produce solutions with good proximity and
diversity.
Eren and Gner [63,64] presented an integer programming model, five TS methods, one random search method,
and a heuristic for the problem of minimizing the weighted
sum of total completion time and the makespan. Setup time
and learning effect are respectively discussed in these
papers. Eren and Gner [65] compared the modified NEH
[66], a TS algorithm, and EDD rule to minimize the
weighted sum of total completion time, total tardiness, and
the makespan.
Tavakkoli-Moghaddam et al. [67] first suggested a
hybrid multi-objective immune algorithm (HMOIA) based
on the features of a biological immune system and bacterial
optimization and stated that HMOIA outperforms SPEA2
and NSGA2, etc., and then they [68] presented an effective
multi-objective immune algorithm (MOIA) for the same
problem and showed that MOIA performs better than

932

SPEA2. The objective is to minimize the weighted mean


completion time and the weighted mean tardiness.
The applications of ACO to FSSP were studied by Li
and Zhang [69] and Yagmahan and Yenisey [70] to
minimize the makespan and the total completion time; the
makespan, the total flow time, and the total machine idle
time, respectively. Tkindt et al. [71] suggested an ACObased lexicographical method to minimize the makespan
and the total completion time.
Ruiz and Allahverdi [72] addressed the problem with the
weighted sum of the makespan and the maximum tardiness
subject to a maximum tardiness threshold vale. They
suggested the global and local dominance relationships for
the three-machine problem and a fast and effective GA for
the general m-machine problem. They showed that the
proposed GA outperforms two existing heuristics.
Noorul Haq and Rahda Ramanan [73] demonstrated that
neural network approach has very promising properties for
real world flow shop sequencing problems. Mansouri et al.
[74] presented a MOGA and a MOSA to find approximations of Pareto-optimal sets for the problem involving
number of setups and the makespan. Liao et al. [75] solved
two bi-criteria problems using a branch-and-bound approach.
Wei et al. [76] presented an effective EA for hybrid flow
shop scheduling in which matrix code is used and the
fitness of each individual is determined with a selective
weight. Computational results show that the proposed EA is
more effective than some existing algorithms.

5 Deterministic scheduling: job shop


Job shop scheduling problem (JSSP) is one of the most
typical and complicated problems. The aim of multiobjective job shop scheduling problem is to allocate n jobs
on m machines in order to optimize multiple performance
indices. This section addresses multi-objective scheduling
literature in job shop environment, and Table 3 shows the
summary of the literature.
Unlike FSSP, JSSP is seldom solved using the conventional techniques for the high complexity of the problem and
the limited optimization ability of the conventional methods.
Meta-heuristics such as GA and EA have become the main
approaches to solve JSSP and flexible job shop scheduling
problem (fJSSP); moreover, the most of flexible scheduling
problems have three objectives for the extra consideration of
workload of machines. Only two papers considered the
special constraints including setup time and batch.
5.1 Job shop
Low et al. [77] considered the problem with reentrant
operations and the dependent setup times. They first

Int J Adv Manuf Technol (2009) 43:926938

constructed three optimization models where each model


investigates a single scheduling objective of minimizing the
total flow time, the total tardiness, and machine idle time,
respectively. They then used a global criterion method to
find a trade-off schedule while considering three objectives
simultaneously.
Ponnambalam et al. [78] proposed a MOGA to derive the
optimal machine-wise priority dispatching rules to resolve
the conflict among the contending jobs in the Giffler and
Thompson (GT) procedure [79]. Esquivel et al. [80]
suggested an enhanced EA with new multi-recombinative
operators and incest prevention strategy for single- and
multi-objective job shop scheduling problem.
Lei and Wu [81, 82] developed a crowding measurebased MOEA which makes use of crowding measure to
adjust the external archive and assign a fitness for each
individual. They showed that the proposed MOEA has
better performance than SPEA.
Chiang and Fu [83] presented a cyclic fitness assignment
mechanism by combining fitness assignment mechanism of
MOGA [19], SPEA2, and NSGA2 and designed a rulecoded GA to minimize the total tardiness and the maximum
tardiness. Petrovic et al. [84] developed a decision support
tool that combines GAs and fuzzy sets. The tool is an
interactive one that allows the decision maker to set
aspiration level and modify them sequentially.
Suresh and Mohanasundaram [85] studied the problem
with objectives of the makespan and the mean flow time of
jobs and suggested Pareto archived simulated annealing
(PASA) which uses Pareto dominance and a simple
aggregating function to accept the candidate solution
among the neighborhood set of solutions. The performance
of PASA is shown to be better than the modified MOGLS
[25] and MOSA [31].
Lei [86] first converted the problem with the objectives
of the makespan and the total tardiness into a continuous
optimization problem and then presented a Pareto archive
particle swarm optimization in which the global best
position selection is combined with crowding measurebased archive maintenance. He showed that the proposed
algorithm performs better than SPEA2, etc. for most of the
chosen benchmark problems.
Qian et al. [87] suggested a memetic algorithm based on
differential evolution (MODEMA) to minimize the makespan and the total tardiness, which applies simultaneously
the DE-based global search and an adaptive local search to
balance the exploration and exploitation. Computational
results showed that MODEMA tends to find non-dominated
solutions with higher quality than IM-MOGLS.
Kleeman and Lamount [88] discussed that some realworld problems can be modeled as the multi-component
scheduling and presented MOEAs with fixed and variable
length chromosome. For general job shop scheduling

Int J Adv Manuf Technol (2009) 43:926938

933

Table 3 Researches on multi-objective job shop scheduling problems


References

Problem

Objectives

Approach

Special constraints

Low et al. [77]

JSSP

Three

Global criterion Method

Ponnambalam et al. [78]


Esquivel et al. [79]
Lei and Wu [81, 82]
Chiang and Fu [83]
Petrovic et al. [84]
Suresh and Mohanasundaram [85]
Lei [86]
Qian et al. [87]
Kleeman and Lamount [88]
Vilcot and Billaut [89]
Zhu et al. [90]
Baykasolu et al. [91]
Kacem et al. [93]
Gao et al. [94]
Xia and Wu [95]
Xing et al. [96]
Kacem et al. [97]
Wu et al. [98]
Liu et al. [99]
Vicot et al. [100]
Tay and Ho [101]
Loukil et al. [102]
Sankar et al. [103]

JSSP
JSSP
JSSP
JSSP
JSSP
JSSP
JSSP
JSSP
Multi-component scheduling
GJSSP
Cost-based JSSP
fJSSP
fJSSP
fJSSP
fJSSP
fJSSP
fJSSP
fJSSP
fJSSP
fJSSP
fJSSP
fJSSP
FMS

Three
Two
Two
Two
Three
Two
Two
Tw
Two
Tw
Two
Three
Three
Three
Three
Three
Two
Two
Two
Two
Three
Up to four
Two

GA
GA
EA
GA
Fuzzy MOGA
SA
PSO
DE and local search
EA
TS and GA
TS
TS
EA and fuzzy logic
GA and bottle shifting
PSO and SA
Simulation
EA and fuzzy logic
Immune genetic algorithm
PSO
GA
GP
SA
GA

Reentrant operations,
the dependent setup
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
Batch
NO

problem (GJSSP), Vilcot and Billaut [89] suggested two


fast and elitist GAs based on NSGA2 in which the initial
population is either randomly generated or partially
generated by TS. Zhu et al. [90] proposed a TS algorithm
to improve the initial scheduling of the cost-based job shop
problem.
5.2 Flexible job shop
Baykasolu et al. [91] presented a linguistic-based model in
which fJSSP is presented as a grammar and the productions
in the grammar are defined as controls. By using these
controls and GT procedure along with the multi-objective
tabu search of Baykasolu et al. [92], fJSSP with three
objectives is solved.
The problem involving the makespan, the maximal
machine workload, and the total workload has been studied
by Kacem et al. [93] who presented a Pareto approach based
on the hybridization of fuzzy logic and EA. The hybrid
approach uses the knowledge representation capabilities of
fuzzy logic and the adaptive capabilities of EA. They
concluded that the hybrid approach cannot guarantee the
optimality but can provide good solutions in a reasonable
time. Gao et al. [94] suggested a GA- and bottleneckshifting-based hybrid algorithm for the same problem,

which uses two-vector representation method, the advanced


crossover, and mutation operators to adapt to the special
chromosome structure and the characteristics of the
problem. They showed that the hybrid algorithm is very
competitive with other methods from literature.
For the problem involving the makespan, the total
workload, and the workload of the critical machine, Xia
and Wu [95] proposed a PSO- and SA-based hybrid
algorithm which makes use of PSO to assign operations
on machines and SA to arrange operations on each
machine. Xing et al. [96] developed a simulation model
for the same problem. After the simulation model is
created, some ACOs and local search-based improvements
are done for the simulation model.
Kacem et al. [97] studied the problem involving the
makespan and the balancing of workload of machines and
suggested an efficient methodology, the first step of which
is to apply the approach by localization to allocate resource
and the second step of which is to apply a controlled
evolutionary approach. They showed that the proposed
method can produce solutions with good quality in a
reasonable computation limit.
Wu et al. [98] proposed a multi-objective immune
genetic algorithm by integrating immune algorithm and
GA. Liu et al. [99] presented a hybrid algorithm based on

934

Int J Adv Manuf Technol (2009) 43:926938

variable neighborhood search and PSO. Vicot et al. [100]


studied the problem of minimizing the makespan and the
maximum tardiness by using a GA for an approximation of
the Pareto optimal front.
Tay and Ho [101] evaluated and employed suitable
parameter and operator spaces for evolving composite
dispatching rules using genetic programming (GP). Experimental results show that the composite dispatching rules
generated by GP outperform the single dispatching rules
and the composite dispatching rules selected from literature
over five large validation sets with respect to the makespan,
the mean tardiness, and the mean flow time.
Loukil et al. [102] dealt with a production scheduling
problem in flexible job shop with particular constraints:
batch production. They presented a MOSA to simultaneously minimize different objectives among the makespan,
the mean completion time, the maximum tardiness, and the
mean tardiness. Sankar et al. [103] suggested a GA with
two different coding schemes for flexible manufacturing
systems with bi-criteria: the total penalty cost and machine
utilization.

6 Uncertain scheduling problems


Manufacturing systems are often subject to some uncertain
events which may disturb their working process: machine
failure, operator unavailability, out-of-stock condition,
changes in availability date, and the latest completion time.
It is realistic to consider a system in an uncertain context;
however, the research on multi-objective scheduling under
uncertainty is still in infancy. Table 4 summarizes the
results for uncertain scheduling problems involving multiple objectives.
Sakawa and Kubota [104] considered JSSP with fuzzy
processing time and fuzzy due date and presented a GA
incorporating the concept of similarity among individuals
using Gantt charts. The objective is to maximize the
minimum agreement index and the average agreement
index and to minimize the maximum fuzzy completion
time. Li et al. [105] proposed a GA for the fuzzy problem

with alternative machines by adopting two-chromosome


presentation and the extended GT Procedure.
Lei [106] addressed the fuzzy problem with objectives of
the minimum agreement index, the maximum fuzzy
completion time, and the mean fuzzy completion time. He
presented an efficient Pareto archive particle swarm
optimization in which the global best position selection is
combined with the crowding measure-based archive maintenance. Xing et al. [107] presented a MOGA for fuzzy
scheduling problem with objectives of the minimum
agreement index and the average agreement index. Ghrayeb
[108] presented a bi-criteria genetic algorithm to minimize the integral value and the uncertainty of the fuzzy
makespan.
Javadi et al. [109] developed a fuzzy multi-objective
linear programming model for multi-objective no-wait flow
shop scheduling in a fuzzy environment. The proposed
model attempts to simultaneously minimize the weighted
mean completion time and the weighted mean earliness.
Lei and Xiong [110] addressed stochastic job shop
scheduling problems (SJSSP) in which the processing time
is modeled by a random variable. They first presented a
permutation-based representation method and then
designed an efficient MOEA to minimize the expected
makespan and the expected total tardiness

7 Summary and conclusions


This paper provided an extensive review of more than 90
papers on MOSP. The methods involved are EA, GA, TS,
SA, GP, ACO, DE, heuristic, local search, and programming method, etc. About these methods, EA and GA are
adopted in more than 40 papers, TS is used in eight papers,
SA in 7seven papers, ACO in six papers, and PSO in six
papers. Immune algorithm is considered in three papers and
GP is adopted in only one paper. The hybrid algorithm was
used in about 20 papers. Some algorithms are seldom
applied to solve some problems; for instance, ACO and SA
are hardly used to handle single machine scheduling,
parallel machine scheduling, fuzzy scheduling, and sto-

Table 4 Researches on multi-objective scheduling under uncertainty


References

Problem

Objectives

Approach

Special constraints

Sakawa and Kubota [104]


Li et al. [105]
Lei [106]
Javadi et al. [107]
Xing et al. [108]
Ghrayeb [109]
Lei and Xiong [110]

FJSSP
FJSSP
FJSSP
FFSSP
FJSSP
FJSSP
SJSSP

Three
Three
Three
Two
Two
Two
Two

GA
GA
PSO
Fuzzy multi-objective linear programming
GA
GA
EA

NO
Alternative process plan
NO
NO
NO
NO
NO

Int J Adv Manuf Technol (2009) 43:926938

chastic scheduling. Thus, hybrid algorithms and metaheuristics except EA and GA should be investigated fully
to find more effective paths to obtain high-quality solutions
in the near future.
Most of the published papers considered the deterministic problem. Few papers addressed fuzzy scheduling problem, and only one paper studied stochastic
scheduling problem. Since most of real life scheduling
problems involves uncertainty and multiple objectives,
future researches on multi-objective scheduling with
fuzzy or stochastic processing conditions are desirable
and attractive.
Only 14 papers considered the special processing
constraints such as setup time, breakdown, and limited
buffer, etc. Most of papers neglected these practical
processing constraints and just focused on the classical
FSSP, JSSP, and FJSSP, etc. In fact, there are a variety of
processing constraints, and some of them are not investigated fully even in a single objective context. Thus, multiobjective scheduling problems with at least one kind of
processing constraint should attract much attention in the
near future.
More than 70 papers considered bi-objective scheduling
problems, and only three papers presented the problem with
more than three objectives, called many-objective scheduling problem. The methods of two- or three-objective
scheduling may have low optimization ability to handle
many-objective problem. The many-objective scheduling
problems pose new challenges for scheduling algorithm
design and implementation.
More than 50 papers dealt with flow shop scheduling,
only 12 papers considered flexible job shop scheduling, and
one paper discussed hybrid flow shop scheduling problem.
Flexible (hybrid) scheduling problem with multiple objectives should be a worthy topic of research. In general,
flexible scheduling problem itself is more complex than the
one with no consideration of flexible process plan. With the
increasing of objectives, the complication of the problem
increases notably. Thus, the efficiency of the scheduling
approach is vital to solve the problem.
Acknowledgments This research is supported by China Hubei
Provincial Science and Technology Department under grant science
foundation project (2007ABA332). The authors also want to express
their deepest gratitude to the anonymous reviewers for their incisive
and seasoned suggestions.

References
1. Nagar A, Haddock J, Heragu S (1995) Multiple and bicriteria
scheduling: a literature survey. Eur J Oper Res 81:88104
doi:10.1016/0377-2217(93)E0140-S
2. TKindt V, Billaut J, Proust C (2001) Multicriteria scheduling
problems: a survey. RAIRO Oper Res 35:143163

935
3. Hoogeveen H (2005) Multicriteria scheduling. Eur J Oper Res
167:592623 doi:10.1016/j.ejor.2004.07.011
4. Knowles JD, Corne DW (2000) Approximating the nondominated front using the Pareto archive evolutionary strategy.
Evol Comput 8(2):149172 doi:10.1162/106365600568167
5. Zitzler E, Thiele L (1999) Multi-objective evolutionary algorithms: a comparative case study and the strength Pareto
approach. IEEE Trans Evol Comput 3(4):257271 doi:10.1109/
4235.797969
6. Zitzler E, Laumanns M, Thiele L (2001) SPEA2: improving the
strength Pareto evolutionary algorithm Tech-Rep. Swiss Federal
Institute of Technology, Lausanne, Switzerland, p 103
7. Kksalan M, Keha AB (2003) Using genetic algorithms for
single-machine bicriteria scheduling problems. Eur J Oper Res
145:543556 doi:10.1016/S0377-2217(02)00220-5
8. Gupta AK, Sivakumar AI (2005) Single machine scheduling with
multiple objectives in semiconductor manufacturing. Int J Adv
Manuf Technol 26:950958 doi:10.1007/s00170-004-2074-y
9. Azizoglu M, Kondakci S, Kksalan M (2003) Single machine
scheduling with maximum earliness and number tardy. Comput
Ind Eng 45:257268 doi:10.1016/S0360-8352(03)00034-2
10. Jolai F, Rabbani M, Amalnick S, Dabaghi A, Dehghan M, Parast
MY (2007) Genetic algorithm for bi-criteria single machine
scheduling problem of minimizing maximum earliness and number
of tardy jobs. Appl Math Comput 194:552560 doi:10.1016/j.
amc.2007.04.063
11. Haral U, Chen R-W, Ferrell WGJ, Kurz MB (2007) Multiobjective
single machine scheduling with non-traditional requirements. Int J
Prod Econ 106:574484 doi:10.1016/j.ijpe.2006.06.018
12. Eren T, Gner E (2006) A bicriteria scheduling with sequencedependent setup times. Appl Math Comput 179:378385
doi:10.1016/j.amc.2005.11.112
13. Chen W-J (2007) An efficient algorithm for scheduling jobs on a
machine with periodic maintenance. Int J Adv Manuf Technol
34:11731182 doi:10.1007/s00170-006-0689-x
14. Suresh V, Chaudhuri D (1996) Bicriteria scheduling problem for
unrelated parallel machines. Comput Ind Eng 30:7782
doi:10.1016/0360-8352(95)00028-3
15. Tkindt V, Billaut J-C, Proust C (2001) Solving a bicriteria
scheduling on unrelated parallel machines occurring in the glass
bottle industry. Eur J Oper Res 135:4249 doi:10.1016/S03772217(00)00288-5
16. Cochran JK, Horng S-M, Fowler JW (2003) A multi-population
genetic algorithm to solve multi-objective scheduling problems
for parallel machines. Comput Oper Res 30:10871102
doi:10.1016/S0305-0548(02)00059-X
17. Murata T, Ishibuchi H, Tanaka H (1996) Multi-objective genetic
algorithm and its applications to flow shop scheduling. Comput
Ind Eng 30:957968 doi:10.1016/0360-8352(96)00045-9
18. Chang P-C, Chen S-H, Lin K-L (2005) Two-phase sub-population
genetic algorithm for parallel- machine scheduling problem.
Expert Syst Appl 29:705712 doi:10.1016/j.eswa.2005.04.033
19. Fonseca CM, Fleming PJ (1993) Genetic algorithms for multiobjective optimization: formulation, discussion and generalization. Proceedings of the Fifth international Conference on
Genetic algorithms. San Mateo, California, pp 416423
20. Deb K, Pratap A, Agarwal S, Meyarivan T (2002) A fast and
elitist multi-objective genetic algorithms: NSGA2. IEEE Trans
Evol Comput 6(2):182197 doi:10.1109/4235.996017
21. Chang P-C, Chen S-H, Hsieh J-C (2006) A global archive subpopulation genetic algorithm with adaptive strategy in multiobjective parallel-machine scheduling problem. Proceedings of
International Conference on Natural Computation, pp 730739
22. Eren T (2008) A bicriteria parallel machine scheduling with
a learning effect of setup and removal times. Appl Math
Model (in press)

936
23. Danneberg D, Tautenhahn T, Werner F (1999) A comparison of
heuristic algorithms for flow shop scheduling problems with
setup times and limited batch size. Math Comput Model 29:101
126 doi:10.1016/S0895-7177(99)00085-0
24. Sivrikaya-Serifolu F, Ulusoy G (1998) A bicriteria two-machine
permutation flowshop problem. Eur J Oper Res 107:414430
doi:10.1016/S0377-2217(97)00338-X
25. Ishibuchi H, Yoshida T, Murata T (2003) Balance between
genetic search and local search in memetic algorithms for multiobjective permutation flow shop. IEEE Trans Evol Comput 17
(2):204223 doi:10.1109/TEVC.2003.810752
26. Geiger MJ (2007) On operators and search space topology in
multi-objective flow shop scheduling. Eur J Oper Res 181:195
206 doi:10.1016/j.ejor.2006.06.010
27. Rajendran C, Ziegler H (2004) Ant-colony algorithms for
permutation flow shop scheduling to minimize makespan/total
flowtime of jobs. Eur J Oper Res 155:426438 doi:10.1016/
S0377-2217(02)00908-6
28. Stuetzle T (1998) An ant approach for the flow shop problem.
Proceedings of European Congress on Intelligent Techniques
and Soft Computing (EUFIT98). Verlag Mainz, Aachen,
Germany, pp 15601564
29. Pasia JM, Hartl RF, Doerner KF (2006) Solving a bi-objective
flowshop scheduling problem by Pareto-ant colony optimization.
Proceedings of M. Dorigo et al. (eds.) ANTS, pp 294305
30. Loukil T, Teghem J, Tuyttens D (2005) Solving multi-objective
production scheduling problems using metaheuristics. Eur J Oper
Res 161:4261 doi:10.1016/j.ejor.2003.08.029
31. Varadharajan TK, Rajendran C (2005) A multi-objective simulated-annealing algorithm for scheduling in flowshops to
minimize the makespan and total flowtime of jobs. Eur J Oper
Res 167:772795 doi:10.1016/j.ejor.2004.07.020
32. Armentano VA, Arroyo JEC (2004) An application of a multiobjective tabu search algorithm to a bi-criteria flow shop problem.
J Heur 10:463481 doi:10.1023/B:HEUR.0000045320.79875.e3
33. Basseur M (2006) Design of cooperative algorithms for multiobjective optimization: application to the flow shop scheduling
problem. Quart J Oper Res 4:255258
34. Pasupathy T, Rajendran C, Suresh RK (2006) A multi-objective
genetic algorithm for searching in flow shop to minimize the
makespan and the total flow time of jobs. Int J Adv Manuf
Technol 27:804815 doi:10.1007/s00170-004-2249-6
35. Rahimi-Vahed AR, Mirghorbani SM (2007) A multi-objective
particle swarm for a flow shop scheduling problem. J Comb
Optim 13:79102 doi:10.1007/s10878-006-9015-7
36. Guo WZ, Chen GL, Min H, Chen S (2007) A discrete particle
swarm optimization algorithm for the multi-objective permutation flow shop sequencing problem. Proceeding of International
Conference on Fuzzy Information and Engineering, pp 323331
37. Qian B, Wang L, Huang D-X, Wang W-L, Wang X (2009) An
effective hybrid DE-based algorithm for multi-objective flow shop
scheduling with limited buffers. Comput Oper Res 36:209233
38. Allahverdi A (2003) The two- and m-machine flowshop
scheduling problems with bicriteria of makespan and mean
flowtime. Eur J Oper Res 147:373396 doi:10.1016/S0377-2217
(02)00253-9
39. Allahverdi A (2004) A new heuristic for m-machine flowshop
scheduling problem with bicriteria of makespan and maximum
tardiness. Comput Oper Res 31:157180 doi:10.1016/S03050548(02)00143-0
40. Allahverdi A, Aldowaisan T (2004) No-wait flowshops with
bicriteria of makespan and maximum lateness. Eur J Oper Res
152:132147 doi:10.1016/S0377-2217(02)00646-X
41. Allahverdi A, Aldowaisan T (2002) No-wait flowshops with
bicriteria of makespan and total completion time. J Oper Res Soc
53:10041015 doi:10.1057/palgrave.jors.2601403

Int J Adv Manuf Technol (2009) 43:926938


42. Allahverdi A, Mittenthal J (1998) Dual criteria scheduling on a
two-machine flowshop subject to random breakdowns. Int Trans
Oper Res 5:317324 doi:10.1016/S0969-6016(97)00042-7
43. Allahverdi A (2001) The tricriteria two-machine flowshop
scheduling problem. Int Trans Oper Res 8:403425 doi:10.1111/
1475-3995.00273
44. Allahverdi A, Savsar M (2002) Scheduling on three-serial
duplicate stations in assembly lines with multiple criteria. Int J
Ind Eng 9:265274
45. Allahverdi A, Al-Anzi FS (2008) The two-stage assembly
flowshop scheduling problem with bicriteria of makespan and
mean completion time. Int J Adv Manuf Technol 37:166177
doi:10.1007/s00170-007-0950-y
46. Yeh WC, Allahverdi A (2004) A branch-and-bound algorithm
for the three-machine flowshop scheduling problem with
bicriteria of makespan and total flowtime. Int Trans Oper Res
11:341359 doi:10.1111/j.1475-3995.2004.00461.x
47. Rajendran C (1995) Heuristics for scheduling in flow shop with
multiple objectives. Eur J Oper Res 82:540555 doi:10.1016/
0377-2217(93)E0212-G
48. Ravindran D, Noorul Haq A, Selvakuar SJ, Sivaraman R (2005)
Flow shop scheduling with multiple objective. Int J Adv Manuf
Technol 25:10071012 doi:10.1007/s00170-003-1926-1
49. Chou F-D, Lee C-F (1999) Two-machine flowshop scheduling
with bicriteria problem. Comput Ind Eng 36:549564
doi:10.1016/S0360-8352(99)00149-7
50. Nagar A, Haddock J, Heragu S (1996) A combined branch-andbound and genetic algorithm based approach for flow shop
scheduling problem. Ann Oper Res 63:397414 doi:10.1007/
BF02125405
51. Basseur M, Lemesre J, Dhaenens C, Talbi E-G (2004)
Cooperation between branch and bound and evolutionary
approaches to solve a bi-objective flow shop problem. Proceedings of Third International Workshop on Experimental and
Efficient Algorithms, pp 7286
52. Sayin S, Karabati S (1999) A bicriteria approach to the twomachine flow shop scheduling problem. Eur J Oper Res
113:435449 doi:10.1016/S0377-2217(98)00009-5
53. Toktas B, Azizoglu M, Koksalan SK (2004) Two-machine flow
shop scheduling with two criteria: maximum earliness and makespan. Eur J Oper Res 157:286295 doi:10.1016/S0377-2217(03)
00192-9
54. Neppalli VR, Chen C-L, Gupta JND (1996) Genetic algorithms
for the two-stage bicriteria flowshop problem. Eur J Oper Res
95:356373 doi:10.1016/0377-2217(95)00275-8
55. Sridhar J, Rajendran C (1996) Scheduling in flow shop and
cellular manufacturing systems with multiple objectivesa
genetic algorithmic approach. Prod Plann Contr 7:374382
doi:10.1080/09537289608930365
56. Chang P-C, Chen S-H, Liu C-H (2007) Sub-population genetic
algorithm with mining gene structures for multi-objective flow
shop scheduling problems. Expert Syst Appl 33:762771
doi:10.1016/j.eswa.2006.06.019
57. Ishibuchi H, Murata T (1998) A multi-objective genetic local search
and its application to flow shop scheduling. IEEE Trans Syst Man
Cybern Part C 28(3):392403 doi:10.1109/5326.704576
58. Talbi EG, Rahoual M, Mabed MH, Dhaenens C (2001) A hybrid
evolutionary approach for multi-criteria optimization problems:
application to flow shop. Proceedings of the First International
Conference on Evolutionary Multi-Criterion Optimization.
Springer, pp 416428
59. Arroyo JEC, Armentano VA (2005) Genetic local search for
multi-objective flow shop scheduling problems. Eur J Oper Res
167:717738 doi:10.1016/j.ejor.2004.07.017
60. Brizuela C, Sannomiya N, Zhao Y (2001) Multi-objective flow
shop: preliminary results. Proceedings of the First International

Int J Adv Manuf Technol (2009) 43:926938

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.
76.

77.

78.

79.
80.

Conference on Evolutionary Multi-Criterion Optimization.


Springer, pp 443457
Ponnambalam SG, Jagannathan H, Kataria M, Gadicherla A
(2004) A TSP-GA multi-objective algorithm for flow-shop
scheduling. Int J Adv Manuf Technol 23:909915 doi:10.1007/
s00170-003-1731-x
Li B-B, Wang L (2007) A hybrid quantum-inspired genetic algorithm
for multi-objective flow shop scheduling. IEEE Trans Syst Man
Cybern Part B. 37:576591 doi:10.1109/TSMCB.2006.887946
Eren T, Gner E (2006) A bicriteria flowshop scheduling
problem with setup time. Appl Math Comput 183:12921300
doi:10.1016/j.amc.2006.05.160
Eren T, Gner E (2008) A bicriteria flowshop scheduling with a
learning effect. Appl Math Model 32(9):17191733 doi:10.1016/
j.apm.2007.06.009
Eren T, Gner E (2008) The triceiteria flowshop scheduling
problem. Int J Adv Manuf Technol 36:12101220 doi:10.1007/
s00170-007-0931-1
Nawaz M, Enscore EE, Ham I (1983) A heuristic algorithm for
the m-machine, n-job flow shop sequencing problem. Omega
11:9195 doi:10.1016/0305-0483(83)90088-9
Tavakkoli-Moghaddam R, Rahimi-Vahed A, Mirzaei AH (2007)
A hybrid multi-objective immune algorithm for a flow shop
scheduling problem with bi-objectives: weighted mean completion time and weighted mean tardiness. Inf Sci 177:50725090
doi:10.1016/j.ins.2007.06.001
Tavakkoli-Moghaddam R, Rahimi-Vahed A, Mirzaei AH (2008)
Solving a multi-objective no-wait flow shop scheduling problem
with an immune algorithm. Int J Adv Manuf Technol 36:968
981 doi:10.1007/s00170-006-0906-7
Li JF, Zhang W (2006) Solution to multi-objective optimization
of flow shop problem based on ACO algorithm. Proceedings of
International Conference on Computational Intelligence and
Security, pp 417420
Yagmahan B, Yenisey MM (2008) Ant colony optimization for
multi-objective flow shop scheduling problem. Comput Ind Eng
54:411420 doi:10.1016/j.cie.2007.08.003
Tkindt V, Monmarch N, Tercinet F, Lagt D (2002) An ant
colony optimization algorithm to solve a 2-machine bicriteria
flow shop scheduling problem. Eur J Oper Res 142:250257
doi:10.1016/S0377-2217(02)00265-5
Ruiz R, Allahverdi A (2008) Minimizing the bicriteria of
makespan and maximum tardiness with an upper bound on
maximum tardiness. Comput Oper Res (in press)
Noorul Haq A, Radha Ramanan T (2006) A bicriterian flow shop
scheduling using artificial neural network. Int J Adv Manuf
Technol 30:11321138 doi:10.1007/s00170-005-0135-5
Mansouri SA, Hendizadeh SH Bicriteria scheduling of a twomachine flow shop with sequence-dependent setup times. Int J
Adv Manuf Technol doi:10.1007/s00170-008-1439-z
Liao CJ, Yu CJ, Joe CB (1997) Bicriterion scheduling in the twomachine flowshop. J Oper Res Soc 48:929935
Wei Z, Xu X-F, Deng S-C (2006) Evolutionary algorithm for
solving multi-objective flow shop scheduling problem. Comput
Integrated Manuf Syst 12:12271234 (in Chinese)
Low C, Wu T-H, Hsu C-M (2005) Mathematical modeling of
multi-objective job shop scheduling with dependent setups and
re-entrant operations. Int J Adv Manuf Technol 27:181189
doi:10.1007/s00170-004-2137-0
Ponnambalam SG, Ramkumar V, Jawahar N (2001) A multiobjective genetic algorithm for job shop scheduling. Prod Plann
Control 12(8):764774 doi:10.1080/09537280110040424
Giffler B, Thompson GL (1960) Algorithm for solving production scheduling problems. Oper Res 8:487503
Esquivel S, Ferrero S, Gallard R, Salto C, Alfonso H, Schtz M
(2002) Enhanced evolutionary algorithm for single and multi-

937

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

objective optimization in job shop scheduling problem. Knowl


Syst 15:1325 doi:10.1016/S0950-7051(01)00117-4
Lei D, Wu Z (2005) Efficient multi-objective evolutionary
algorithm for job shop scheduling. Chin J Mech Eng 18
(4):494497
Lei D, Wu Z (2006) Crowding-measure-based multi-objective
evolutionary for job shop scheduling. Int J Adv Manuf Technol
30:112117 doi:10.1007/s00170-005-0029-6
Chiang T-C, Fu L-C (2006) Multiobjective job shop scheduling using genetic algorithm with cyclic fitness assignment.
Proceedings of 2006 IEEE Congress on Evolutionary Computation, pp 1621
Petrovic D, Duenas A, Petrovic S (2007) Decision support tool
for multi-objective job shop scheduling problems with linguistically quantified decision functions. Decis Support Syst
43:15271538 doi:10.1016/j.dss.2006.06.006
Suresh RK, Mohanasundaram KM (2006) Pareto archived
simulated annealing for job shop scheduling with multiple
objectives. Int J Adv Manuf Technol 29:184196 doi:10.1007/
s00170-004-2492-x
Lei D (2008) A Pareto archive particle swarm optimization for
multi-objective job shop scheduling. Comput Ind Eng 54:960
971 doi:10.1016/j.cie.2007.11.007
Qian B, Wang L, Huang D-X, Wang X (2008) Scheduling multiobjective job shops using a memetic algorithm based on
differential evolution. Int J Adv Manuf Technol 35:10141027
doi:10.1007/s00170-006-0787-9
Kleeman MP, Lamont GB (2007) Scheduling of flow-shop, jobshop and combined scheduling problems using MOEAs with
fixed and variable length chromosomes. Stud Comput Intell
49:4999 doi:10.1007/978-3-540-48584-1_3
Vilcot G, Billaut J-C (2008) A tabu search and a genetic algorithm
for solving bicriteria general job shop scheduling problem. Eur J
Oper Res 190:398411 doi:10.1016/j.ejor.2007.06.039
Zhu ZC, Ng KM, Ong HL (2007) An application of tabu search
algorithm on cost-based job shop problem with multiple
objectives. Proceeding of IEEE International Conference on
Industrial Engineering and Engineering Management, pp 912
916
Baykasolu A, zbakir L, Snmez A (2004) Using multiple
objective tabu search and grammars to model and solve multiobjective flexible job shop scheduling problems. J Intell Manuf
15:777785 doi:10.1023/B:JIMS.0000042663.16199.84
Baykasolu A, Owen S, Gindy N (1999) A taboo search based approach
to find the Pareto optimal set in multiple objective optimization. J Eng
Optim 31:731748 doi:10.1080/03052159908941394
Kacem I, Hammadi S, Borne P (2002a) Approach by localization
and multi-objective evolutionary optimization for flexible job
shop scheduling problems. IEEE Trans Syst Man Cybern PART
C 32(1):113 doi:10.1109/TSMCC.2002.1009117
Gao J, Gen M, Sun L, Zhao X (2007) A hybrid of genetic
algorithm and bottleneck shifting for multi-objective flexible job
shop scheduling problems. Comput Ind Eng 53:149162
doi:10.1016/j.cie.2007.04.010
Xia W, Wu Z (2005) An effective hybrid optimization approach
for multi-objective flexible job-shop scheduling problems.
Comput Ind Eng 48:409425 doi:10.1016/j.cie.2005.01.018
Xing L-N, Chen Y-W, Yang K-W (2008) Multi-objective flexible
job shop schedule: design and evaluation by simulation
modeling. Appl Soft Comput (in press)
Kacem I, Hammadi S, Borne P (2002b) Pareto-Optimality
approach for flexible job-shop scheduling problems: hybridization of evolutionary algorithms and fuzzy logic. Math Comput
Simul 60:245276 doi:10.1016/S0378-4754(02)00019-8
Wu XL, Sun SD, Niu GG, Zhai YN (2006) The performance
analysis of a multi-objective immune genetic algorithm for

938

99.

100.

101.

102.

103.

104.

Int J Adv Manuf Technol (2009) 43:926938


flexible job shop scheduling. International Federation for
Information Processing, pp 914919
Liu H, Abraham A, Choi O, Moon SH (2006) Variable
neighborhood particle swarm optimization for multi-objective
flexible job-shop scheduling problems. Proceedings of SEAL,
pp 197204
Vicot G, Billaut J-C, Esswein C (2006) A genetic algorithm for a
bicriteria flexible job shop scheduling problem. 2006 International Conference on Service Systems and Service Management,
pp 12401244
Tay JC, Ho NB (2008) Evolving dispatching rules using genetic
programming for solving multi-objective flexible job-shop
problems. Comput Ind Eng 54:453473 doi:10.1016/j.
cie.2007.08.008
Loukil T, Teghem J, Fortemps P (2007) A multi-objective
production scheduling case study solved by simulated annealing.
Eur J Oper Res 179:709722 doi:10.1016/j.ejor.2005.03.073
Sankar SS, Ponnanbalam SG (2003) A multiobjective genetic
algorithm for scheduling a flexible manufacturing systems. Int J
Adv Manuf Technol 22:229236 doi:10.1007/s00170-002-1464-2
Sakawa M, Kubota R (2000) Fuzzy programming for multiobjective job shop scheduling with fuzzy processing time and
fuzzy due date through genetic algorithm. Eur J Oper Res
120:393407 doi:10.1016/S0377-2217(99)00094-6

105. Li FM, Zhu YL, Yin CW, Song XY (2005) Fuzzy programming
for multi-objective fuzzy job shop scheduling with alternative
machines through genetic algorithm. In: Wang L, Chen K, Ong
YS (eds) Advance in natural computation. Springer, Berlin, pp
9921004
106. Lei DM (2008) Pareto archive particle swarm optimization for
multi-objective fuzzy job shop scheduling problems. Int J Adv
Manuf Technol 37:157165 doi:10.1007/s00170-007-0945-8
107. Javadi B, Saidi-Mehrabad M, Haji A, Mahdavi I, Jolai F,
Mahdavi-Amiri N (2008) No-wait flow shop scheduling
using fuzzy multi-objective linear programming. J Franklin
Inst 345(5):452467
108. Xing YJ, Wang ZQ, Sun J, Meng JJ (2006) A multi-objective
fuzzy genetic algorithm for job-shop scheduling problems. 2006
International Conference on Computational Intelligence and
Security, pp 398401
109. Ghrayeb OA (2003) A bi-criteria optimization: minimizing the
integral value and spread of the fuzzy makespan of job shop
scheduling problems. Appl Soft Comput 2:197210 doi:10.1016/
S1568-4946(02)00069-8
110. Lei DM, Xiong HJ (2007) An efficient evolutionary algorithm
for multi-objective stochastic job shop scheduling. Sixth International Conference on Machine Learning and Cybernetics, pp
867872

También podría gustarte