Professional Documents
Culture Documents
To cite this article: Hamid Yilmaz & Mustafa Yilmaz (2019): A mathematical model and tabu
search algorithm for multi-manned assembly line balancing problems with assignment restrictions,
Engineering Optimization, DOI: 10.1080/0305215X.2019.1618288
Article views: 35
1. Introduction
Assembly lines are the most commonly used systems for high-volume production. Assembly lines
include a series of stations arranged according to the material handling method. Different tasks are
handled depending on precedence relations for a given cycle time. The problem of assigning tasks
to workers located on workstations in order to optimize one or more specific objective, such as
minimizing the cycle time for a known number of workstations, minimizing the open workstations
for a known cycle time and maximizing the efficiency of the assembly line subject to the relation-
ships between tasks, is called the assembly line balancing problem (ALBP) (Becker and Scholl 2006).
Numerous researchers have tried to present realistic models and generalized ALBPs. Hence, the litera-
ture contains additional characteristics such as cost minimization, mixed-model production, parallel
assembly lines and multi-manned assembly lines, among others. More detailed reviews of such stud-
ies are presented by Becker and Scholl (2006) and Boysen, Fliedner, and Scholl (2007) and, more
recently, by Battaïa and Dolgui (2013) and Sivasankaran and Shahabudeen (2014).
Multi-manned assembly lines are a new type of generalized assembly line. The balancing problems
of these kinds of lines are known as multi-manned assembly line balancing problems (MALBPs).
Several definitions of MALBP exist in the literature (Kellegöz and Toklu 2015; Yilmaz and Yilmaz
2016). One of the most obvious characteristics of the MALBP is that it allows assembly of the tasks
to be performed simultaneously on the same station, as shown in Figure 1.
To the authors’ knowledge, the first study on MALBP in the literature was presented by Dimitriadis
(2006). Dimitriadis proposed a two-level heuristic method based on a modification of Hoffmann’s
(1963) technique, to solve MALBP with the objective of minimizing the total number of workers on
the line and the number of open multi-manned workstations for a given cycle time. According to sim-
ple assembly line worker numbers, computational experiments showed that the heuristic was effective
and could result in an enhancement of the total space utilization. Fattahi, Roshani, and Roshani (2011)
presented the first mixed-integer mathematical programming formulation for MALBP. The aim of
the mathematical model was to minimize the number of workers as the first purpose and the number
of workstations as the second purpose. A heuristic algorithm, based on ant colony optimization, was
proposed to find the solutions of medium- and large-sized test problems. Then, the simulated anneal-
ing metaheuristic algorithm was used to solve the problem while considering the smoothness index,
line efficiency and line length as the performance criteria by Roshani et al. (2013). Kellegoz (2016)
developed a mixed-integer mathematical model for the ALBP with multi-manned workstations. In
that study, a heuristic approach based on simulated annealing was also proposed. Kellegoz and Toklu
(2012) presented an efficient branch-and-bound algorithm for ALBPs with parallel multi-manned
workstations. They also studied the parallel MALBP and developed a mixed-integer mathematical
programming formulation for the problem. An efficient constructive heuristic algorithm based on
priority rules was presented. A genetic algorithm-based solution procedure to improve solutions
found by the constructive heuristic was presented by Kellegöz and Toklu (2015).
In addition to the developments mentioned above, the MALBP has been studied in different areas.
Becker and Scholl (2009) considered a special case of MALBP with variable parallel workplaces.
The authors proposed an exact solution algorithm for the problem based on the branch-and-bound
rule, which is called VWSolver. Cevikcan, Durmusoglu, and Unal (2009) devised a mathematical
model for multi-manned stations in mixed-model assembly lines. Moon, Logendran, and Lee (2009)
referred to multi-manned assembly lines with skilled workers. Yilmaz and Yilmaz (2016) drew atten-
tion to ALBPs in which workers have been assigned to teams in advance owing to the need for
specialized skills or equipment on the line. They used a heuristic assembly line balancing procedure
and developed a mathematical model of the problem. Giglio et al. (2017) presented a new mixed-
integer programming (MIP) formulation to solve the problem with skilled workers. The proposed
formulation was used to solve some experimental problems found in the literature and showed the
effectiveness of the model. Chang and Chang (2010) examined a mixed-model multi-manned assem-
bly line, and proposed a mathematical formulation for this problem with the objective of minimizing
the number of open multi-manned workstations. Kazemi and Sedighi (2013) studied a cost-oriented
MALBP. They proposed a mathematical model formulation with the objective of minimizing the total
cost per production unit, and presented a heuristic method based on the genetic algorithm to solve
real-sized problem instances. Zamzam et al. (2015) developed a hybrid genetic algorithm to solve
the MALBP and defined a new indicator to determine the maximum allowable number of workers
ENGINEERING OPTIMIZATION 3
per station. Roshani and Giglio (2017) proposed a mixed-integer mathematical programming for-
mulation for the MALBP with the objective of minimizing the cycle time. Approaches based on the
simulated annealing algorithm were developed and the results compared between approaches.
The basic aim of the multi-manned assembly line configuration is to reduce the number of workers
as the primary objective and then to try to minimize the number of open multi-manned workstations.
In addition to these goals, multi-manned assembly lines need new constraints and aims to be applied
in real-life industry. One of these aims is to reduce the difference in the workers’ task load. Yilmaz and
Yilmaz (2015) proposed a mathematical formulation for the MALBP which aimed to minimize the
number of workers and workstations, and the workload difference between workers. Thus, the work-
ing time of each worker was balanced in multi-manned workstations by minimizing the differences
between the task loads. Chen, Cheng, and Li (2018) proposed an MIP model for the MALBP under
resource constraints. The objective of the study was to minimize the number of operators, worksta-
tions and resources to obtain optimal line balancing. Naderi, Azab, and Borooshan (2019) proposed a
balancing problem including some specific requirements (the assembly line has five sides and work-
ers can move along these sides). A mixed-integer linear programming model was proposed for the
problem, and then the model was improved with linear relaxation.
A great deal of research on multi-manned assembly lines has included only the cycle time and
precedence restrictions. In practice, there are usually constraints which restrict the assignment of
workpieces or tasks to stations in addition to the classical constraints.
Up to now, some heuristic procedures and a few exact approaches have been developed for solving
multi-manned assembly line problems under different conditions, but there is no solution procedure
available for multi-manned assembly lines with assignment restriction.
In this article, the MALBP is extended by adding different types of assignment restrictions. A
mixed-integer mathematical formulation is proposed and a metaheuristic approach based on the tabu
search algorithm (TSA) is developed for the problem.
The rest of this article is organized as follows. Section 2 describes common assignment restric-
tions on ALBPs. The mathematical model for the problem is presented in Section 3. Section 4 gives
comprehensive explanations of the proposed TSA. Section 5 presents the numerical experiment and
computational results, and conclusions are given in Section 6.
2. Assignment restrictions
Purnomoa, Wee, and Raua (2013) proposed a mathematical model for an ALBP with assignment
restrictions. In that article, the widest but not limited assignment restrictions were listed for the ALBP.
In this study, resource constraints are ignored because such constraints are presented in other studies
(Chen, Cheng, and Li 2018) for multi-manned assembly line problems. In addition to the cycle time
and precedence relationship constraints, the following types of assignment restriction are considered
in the literature:
• Zoning restrictions: These are also called task restrictions. There are two different kinds of zoning
(task) restriction, as seen in Figure 2, namely positive zoning and negative zoning restrictions.
Positive zoning ensures that a set of tasks is assigned to the same workstation while negative zoning
assigns a set of tasks to different workstations on the line. If a set of tasks requires an expensive
resource (positive zoning constraints), they can share the same workstation. If a set of tasks requires
different equipment located in different stations (negative zoning constraints), they cannot share
the same workstation (Dar-El and Rubinovitch 1979; Boysen, Fliedner, and Scholl 2007).
• Distance restrictions: On the production line, a process may need to observe minimum distances
or maximum distances between tasks. This distance is measured in time, space or station positions
(Buxey 1974; Pastor and Corominas 2000). An example of minimum distance, as seen in Figure 3,
might be observed in cases where a colour has to dry before different task(s) can be performed
on the workpiece. An example of maximum distance might be observed when melted metal must
4 H. YILMAZ AND M. YILMAZ
be prevented from cooling down before a specific task is carried out (Scholl, Fliedner, and Boysen
2010).
• Station restrictions: In industrial applications, there are constraints on assigning task(s) to specific
station(s). For example, assembly tasks which need handling equipment or heavy machinery fixed
to a specific station must be performed in that station, as seen in Figure 4 (Scholl, Fliedner, and
Boysen 2010).
• Synchronous task restrictions: These tasks can only be applied on two-sided and multi-manned
assembly lines. If the tasks must be performed simultaneously by different workers on the same
station, these tasks are synchronous tasks, as seen in Figure 5 (Purnomoa, Wee, and Raua 2013).
To the best of the authors’ knowledge, no studies on the task assignment restrictions mentioned
above have been published for MALBPs.
ENGINEERING OPTIMIZATION 5
3. Mathematical formulation
In this section, a mixed-integer mathematical model is developed for the MALBP under assignment
restrictions. In the proposed mathematical model, the following assumptions are considered:
Indices
i, h, g Tasks
j, n Workstations
k Workers
Nws Number of workstation
Parameters
£ Digit number of J
I Set of tasks, I = {1, 2, . . . , i, . . . h, . . . , I}
J Set of workstations, J = {1, 2, . . . , j, . . . Nws}
K Maximum allowable workers per workstation, K = {1, 2, . . . k, . . . , K max }
Is Set of synchronous tasks
I zp Set of positive zoning tasks
I zn Set of negative zoning tasks
J ki Set of station restricted tasks
Dmin Set of tasks with minimum distance
Dmax Set of tasks with maximum distance
tk(i) Task time of ith task
Ct Cycle time
M A very large number
P(i) Set of immediate predecessors of task i
P1(i) Set of all predecessors of task i
S(i) Set of immediate successors of task i
S1(i) Set of all successors of task i
dmin (i,g) Minimum distance between tasks i and g
dmax (i,g) Maximum distance between tasks i and g
Decision variables
x(i,j) 1, if ith task assigned to workstation j; 0, otherwise
t(i,j,k) 1, if ith task assigned to kth worker in jth workstation; 0, otherwise
st(i) Starting time of task i
w(j,k) 1, if kth worker is assigned to jth workstation; 0, otherwise
U(j) 1, if jth multi-manned workstation is open; 0, otherwise
y(i,h) 1, if task i is executed earlier than task h in the sequence of tasks assigned to the same
worker on the same workstation; 0, otherwise
a(i,g) 1, if the workstation difference between tasks i and g is negative; 0, otherwise
xij = 1, ∀i ∈ I (2)
j∈J
ENGINEERING OPTIMIZATION 7
tijk = xij , ∀i ∈ I and j ∈ J (3)
k∈K
j.xhj ≤ j.xij , ∀i ∈ I, h ∈ P(i) (4)
j∈J j∈J
sti − sth + M.(1 − xij ) + M.(1 − xhj ) ≥ tkh , ∀i ∈ I, h ∈ P(i), j∈J (6)
Uj ≥ Uj+1 , j = 1, 2, . . . , (J − 1) (10)
xij − Uj ≥ 0, ∀j ∈ J (11)
i∈I
tijk − ||I||.wjk ≤ 0, ∀j ∈ J, ∀k ∈ K (12)
i∈I
j.xij − j.xgj ≤ M.(1 − aig ) − dmin (i, g), (i, g) ∈ Dmin (18)
j∈J j∈J
j.xij − j.xgj ≤ dmax (i, g), (i, g) ∈ Dmax (19)
j∈J j∈J
j.xij − j.xgj ≥ −dmax (i, g), (i, g) ∈ Dmax (20)
j∈J j∈J
sti ≥ 0, ∀i ∈ I (21)
8 H. YILMAZ AND M. YILMAZ
solve the medium- and large-sized scales of this problem. Two-sided assembly lines share similarities
with MALBP in having multiple workers on each workstation. One of the most suitable heuristics for
this kind of problem is the TSA (Özcan and Toklu 2009). In addition, the TSA is a powerful heuristic
to solve combinatorial optimization problems. This research adopts a modified version of the TSA
approach to present a heuristic approach for large scales of the MALBP with assignment restrictions.
To the authors’ knowledge, this article makes one of the first attempts to present the MALBP with
assignment restrictions and solution procedures (a mathematical model and a heuristic algorithm
based on the TSA).
Figure 6. Flowchart of tabu search algorithm for the multi-manned assembly line balancing problem.
ENGINEERING OPTIMIZATION 11
the first matched multi-manned workstation according to some given rules and restrictions. Then,
the set of assignable tasks is updated and this process continues until all tasks have been assigned to
workstations. For example, as seen in Figure 7, there are five tasks in the precedence diagram. Before
the assignment process, a PL is generated for the tasks randomly. For the first station, the assignable
tasks are the first and second tasks. The highest value of the candidate tasks is 0.65. So, the assignment
process begin with the first task and continues until all tasks have been assigned.
The above solution mechanism presented by Özcan and Toklu (2009) for balancing of two-sided
assembly lines is used in this article for balancing MALBP.
The neighbour solutions determine the way in which a move is chosen leading to the next step.
The total number of solutions in the neighbourhood is calculated as n − 1.
where f max (LE) is the target value of f (LE) and is set to 100, f min (LN) is the target value of f (LN)
and is set to TSN/K, and f min (SI) is the target value of f (SI) and is set to zero. Thus, a balancing
ENGINEERING OPTIMIZATION 13
solution with zero f (SI), the LN equal to TSN/K and 100% level of f (LE) is a perfectly balanced line.
Because of the structure of the formulation, all calculations have similar characteristics and affect the
line balancing similarly. In this algorithm, the first aim is to minimize the LE because its value can
possibly increase to 100, and the SI is affected directly by the first aim. The second aim is to minimize
the number of workstations on the line directly related to LN.
Step 1: Ist = 1.
Step 2: Define AL. (AL = {i | all i ∈ Öi have already been assigned and all restrictions are
fulfilled}); if the step is directed from Step 3 and AL = { }, Ist = Ist − 1, and go to
Step 14.
Step 3: If AL = { }, Ist = Ist + 1 and go to Step 2. If AL = { }, go to Step 4.
Step 4: Sort the PL and take the greatest value of PL(i) from AL.
Step 5: If the selected task is synchronous with another task, select both tasks and go to Step 10.
Otherwise, continue.
Step 6: If BiZh ≤ Zk (BiZh = max{BiZh | h ∈ Öi at the same station}, k ∈ K), go to Step 8.
Step 7: If some BiZh ≤ Zk and some BiZh > Zk (BiZh = max{BiZh | h ∈ Öi at the same station},
k ∈ K), Zk = BiZh (in fact, BiZh > Zk) and go to Step 8.
Step 8: If there is more than one ‘min Zk’ value (n pieces), generate a random number between 0
and 1. If the random number is between 0 and (1/n), take the value of the first Zk. If the
random number is between 1/n and 2/n, take the value of the second Zk. If the random
Figure 9. Flowchart for building a feasible solution for the multi-manned assembly line balancing problem.
14 H. YILMAZ AND M. YILMAZ
number is between ((n − 1)/n) and 1, take the value of the nth min Zk and BZi = min Zk.
Otherwise, BZi = min Zk, and go to Step 9.
Step 9: If BZi + Zi ≤ Ct , assign the ith task in station Ist to the kth worker, BiZi = BZi + Zi, and
go to Step 2. Otherwise, Ist = Ist + 1, and go to Step 2.
Step 10: If BiZh ≤ Zk (BiZh = max {BiZh | h ∈ (Öi, Öi1) and at the same station}, k ∈ K), go to
Step 12.
Step 11: If some BiZh ≤ Zk and some BiZh > Zk (BiZh = max{BiZh | h ∈ (Öi, Öi1) and at the
same station}, k ∈ K), Zk = BiZh (in fact, BiZh > Zk), and go to Step 12.
Step 12: Sort all Zk values, and select the min Zk and the next min Zk1 values, BZi = Zk1, and go
to Step 13.
Step 13: If BZi + max{Zi, Zi1} ≤ Ct , assign the ith task to the kth worker in station Ist, assign the
i + 1 task to the k + 1 worker in station Ist, BiZi = BZi + Zi, BiZi1 = BZi + Zi1, and go to
Step 2. Otherwise, Ist = Ist + 1, and go to Step 2.
Step 14: Calculate the objective function.
presented by Bowman (1960), Jackson (1956), Jaeschke (1964), Mansoor (1964), Merten (1967) and
Mitchell (1957); three medium-size problems, presented by Sawyer (1970), Heskiaoff (1968) and Kil-
bridge and Wester (1961); and three relatively large-sized problems, presented by Tonge (1961) and
Arcus (1963), constitute the data set.
All of the original problems were discussed with only precedence-constrained restrictions in
MALBP. In this study, the synchronous restriction(s), zoning restriction(s), workstation restriction(s)
and distance restriction(s) are added for each test problem. Thus, the proposed problem provides a
more realistic representation of multi-manned assembly line problems. A mathematical model and
TSA are used to improve the production system with these additional realistic constraints. The restric-
tions are listed in Tables 1 and 2. Considering the problem by Bowman (1960) as an example, the
problem has a pair of synchronous tasks (tasks 4 and 5), a pair of zoning positive tasks (tasks 6 and 7)
and a pair of zoning negative tasks (tasks 1 and 6). There is a task with a workstation restriction (task
1 is assigned to workstation 1), a pair of tasks with minimum distance (tasks 7 and 8 have a minimum
Table 3. Results of the proposed algorithm and mathematical model for small-sized problems.
distance of one workstation between them) and a pair of tasks with maximum distance (tasks 1 and
2 have a maximum distance of two workstations between them). n is the total number of tasks in the
test instances.
The parameters of the TSA are set based on Section 4. The best solutions from each experi-
ment are presented. The total number of workers on the line and the number of open multi-manned
workstations numbers are used to compare the proposed TSA and the mathematical model.
The proposed algorithm is programmed using C# programming language and the mathematical
model is solved using Gurobi 6.0 solver. The set of test problems is solved on an Intel Core I5 2.5 GHz
personal computer with 4 GB RAM.
Small-sized test problems are used to compare the performance of the proposed TSA with the
optimal solution of the proposed mathematical model presented in Section 3. The solutions of the
experiments are reported in Table 3, which shows the optimal number of multi-manned workstations
and the optimal number of workers found by the proposed mathematical model and the TSA for
Table 4. Comparison of the proposed algorithm solutions and mixed-integer programming (MIP) model for medium- and large-
sized problems.
Tabu search algorithm MIP
small-sized test problems. As can be seen in Table 3, TSA can find the optimal solutions to the small-
sized problems. Thus, it can be stated that the proposed TSA performs as well as the mathematical
model in minimizing the numbers of workers and open multi-manned workstations.
Finally, five criteria are considered for medium- and large-sized test problems: the number of work-
ers, the number of workstations, the line efficiency (LE), the smoothness index (SI) and the space
utilization factor (Suf). The maximum time limit for the MIP model is set to 3600 CPU-s. Compar-
isons are made in terms of the number of workstations on the line and the length of the assembly line.
These values are compared between the mathematical model solutions and the TSA solutions for the
numbers of workers and open workstations in the MALBP with assignment restrictions. The MALBP
with assignment restrictions is a problem with additional constraints to the MALBP. For this reason,
it is possible that more workstations and/or workers will be needed.
The computation times of the proposed tabu search are found to be between 0.49 and 1503.55 s.
The performance comparison of the proposed TSA and mathematical model in finding the optimal
number of workers on the assembly line is shown in Table 3. This table shows that the proposed TSA
outperforms the mathematical model in some test problems, according to CPU time. In addition,
optimal solutions are found by the TSA on all small-sized test instances. The proposed TSA finds the
same results for 25 subproblems in four test instances.
Medium- and large-sized test problems are reported in Table 4. From Table 4, the number of addi-
tional workers is between 0 and 4 for the TSA solutions. The mathematical model cannot find a
solution for 35.14% of the test problems in Table 4, whereas the proposed algorithm finds feasible solu-
tions for all test problems. Total CPU times for both solution methods are also presented in Table 4. It
can be seen that TSA’s total solution time is 9.16% of the total MIP solution time. Thus, the proposed
TSA obtains a better solution quality in less CPU time. In addition, both the mathematical model
and the TSA find the same solutions for 17 medium-sized test problems, but the TSA finds these
solutions in 947.18 CPU-s. The total solution time for the mathematical model is 59,717 CPU-s. This
means that the TSA generally obtains better solutions. When both methods find the same solution,
the TSA has better solution times than the mathematical model (the TSA CPU time is 1.58% of the
mathematical model CPU time for the same solutions).
6. Conclusion
In this study, a mathematical model formulation and a TSA for solving the MALBP with assign-
ment restrictions to minimize the number of workers on the line and multi-manned workstation is
proposed for the first time. The proposed TSA is compared with proposed mathematical formula-
tion results using test problems in the literature. The results of the computational study on the test
problems indicate that the proposed algorithm performs well. The proposed algorithm finds feasible
solutions for all of the test problems.
This study presents a good starting point for future research. First, an exact solution methodology
such as the branch-and-bound algorithm may be developed for the problem. Secondly, more effec-
tive metaheuristic approaches such as ant colony, simulated annealing or genetic algorithms may be
developed. Thirdly, the problem can be modelled by taking into account several other criteria, such
as load balancing or other real-life problems. Finally, MALBPs should be addressed by considering
different assembly line layouts.
Disclosure statement
No potential conflict of interest was reported by the authors.
References
Arcus, A. L. 1963. “An Analysis of a Computer Method of Sequencing Assembly Line Operations.” Ph.D. Dissertation,
University of California.
18 H. YILMAZ AND M. YILMAZ
Battaïa, O., and A. Dolgui. 2013. “A Taxonomy of Line Balancing Problems and Their Solution Approaches.”
International Journal of Production Economics 142 (2): 259–277.
Becker, C., and A. Scholl. 2006. “A Survey on Problems and Methods in Generalized Assembly Line Balancing.”
European Journal of Operational Research 168 (3): 694–715.
Becker, C., and A. Scholl. 2009. “Balancing Assembly Lines with Variable Parallel Workplaces: Problem Definition and
Effective Solution Procedure.” European Journal of Operational Research 199 (2): 359–374.
Bowman, E. H. 1960. “Assembly Line Balancing by Linear Programming.” Operations Research 8: 385–389.
Boysen, N., M. Fliedner, and A. Scholl. 2007. “A Classification of Assembly Line Balancing Problems.” European Journal
of Operational Research 183 (2): 674–693.
Buxey, G. M. 1974. “Assembly Line Balancing with Multiple Stations.” Management Science 20: 1010–1021.
Cevikcan, E., M. B. Durmusoglu, and M. E. Unal. 2009. “A Team-Oriented Design Methodology for Mixed Model
Assembly Systems.” Computers & Industrial Engineering 56 (2): 576–599.
Chang, H.-J., and T.-M. Chang. 2010. “Simultaneous Perspective-Based Mixed-Model Assembly Line Balancing
Problem.” Tamkang Journal of Science and Engineering 13: 327–336.
Chen, Y.-Y., C.-Y. Cheng, and J.-Y. Li. 2018. “Resource-Constrained Assembly Line Balancing Problems with Multi-
manned Workstations.” Journal of Manufacturing Systems 48: 107–119.
Dar-El, E. M., and Y. Rubinovitch. 1979. “MUST—A Multiple Solutions Technique for Balancing Single Model
Assembly Lines.” Management Science 25: 1105–1114.
Dimitriadis, S. G. 2006. “Assembly Line Balancing and Group Working: A Heuristic Procedure for Workers’ Groups
Operating on the Same Product and Workstation.” Computers and Operations Research 33: 2757–2774.
Fattahi, P., A. Roshani, and A. Roshani. 2011. “A Mathematical Model and Ant Colony Algorithm for Multi-
manned Assembly Line Balancing Problem.” International Journal of Advanced Manufacturing Technology 53:
363–378.
Giglio, D., M. Paolucci, A. Roshani, and F. Tonelli. 2017. “Multi-manned Assembly Line Balancing Problem with Skilled
Workers: A New Mathematical Formulation.” IFAC-PapersOnLine 50 (1): 1211–1216.
Glover, F. 1989. “Tabu Search—Part I.” ORSA Journal on Computing 1: 190–206.
Glover, F. 1990. “Tabu Search—Part II.” ORSA Journal on Computing 2: 4–32.
Heskiaoff, H. 1968. “A Heuristic Method for Balancing Assembly Lines.” Western Electric Engineer 12: 9–16.
Hoffmann, T. R. 1963. “Assembly Line Balancing with a Precedence Matrix.” Management Science 9: 551–562.
Jackson, J. R. 1956. “A Computing Procedure for a Line Balancing Problem.” Management Science 2: 261–271.
Jaeschke, G. 1964. “Eine Allgemaine Methode Zur Losung Kombinatoriiicher Probleme.” Ablauf-Plan Forsch 5:
133–153.
Kazemi, A., and A. Sedighi. 2013a. “A Cost-Oriented Model for Multi-manned Assembly Line Balancing Problem.”
Journal of Optimization in Industrial Engineering 2 (13): 13–25.
Kellegoz, T. 2016. “Assembly Line Balancing Problems with Multi-manned Stations: A New Mathematical Formulation
and Gantt Based Heuristic Method.” Annals of Operations Research 253 (1): 377–404.
Kellegoz, T., and B. Toklu. 2012. “An Efficient Branch and Bound Algorithm for Assembly Line Balancing Problems
with Parallel Multi-manned Workstations.” Computers & Operations Research 39 (12): 3344–3360.
Kellegöz, T., and B. Toklu. 2015. “A Priority Rule Based Constructive Heuristic and an Improvement Method for Bal-
ancing Assembly Lines with Parallel Multi-manned Workstations.” International Journal of Production Research 53:
736–756.
Kilbridge, M. D., and L. Wester. 1961. “A Heuristic Method of Assembly Line Balancing.” Journal of Industrial
Engineering 12: 292–298.
Mansoor, E. M. 1964. “Assembly Line Balancing—An Improvement on the Ranked Positional Weight Technique.”
Journal of Industrial Engineering 15: 73–77.
Merten, P. 1967. “Assembly Line Balancing by Partial Enumeration.” Ablauf- und planungsforschung 8: 429–433.
Mitchell, J. 1957. A Computational Procedure for Balancing Zoned Assembly Lines. Research Report 6-94801-1-R3.
Westinghouse Research Laboratories, Pittsburgh.
Moon, I., R. Logendran, and J. Lee. 2009. “Integrated Assembly Line Balancing with Resource Restrictions.” Interna-
tional Journal of Production Research 47 (19): 5525–5541.
Naderi, B., A. Azab, and K. Borooshan. 2019. “A Realistic Multi-manned Five-Sided Mixed-Model Assembly Line Bal-
ancing and Scheduling Problem with Moving Workers and Limited Workspace.” International Journal of Production
Research 57 (3): 643–661.
Özcan, U., and B. Toklu. 2009. “A Tabu Search Algorithm for Two-Sided Assembly Line Balancing.” International
Journal of Advanced Manufacturing Technology 43: 822–829.
Pastor, R., and A. Corominas. 2000. “Assembly Line Balancing with Incompatibilities and Bounded Workstation Loads.”
Ricerca Operativa 30: 23–45.
Purnomoa, H. D., H. Wee, and H. Raua. 2013. “Two-Sided Assembly Lines Balancing with Assignment Restrictions.”
Mathematical and Computer Modelling 57: 189–199.
Roshani, A., and D. Giglio. 2017. “Simulated Annealing Algorithms for the Multi-manned Assembly Line Balancing
Problem: Minimising Cycle Time.” International Journal of Production Research 55 (10): 2731–2751.
ENGINEERING OPTIMIZATION 19
Roshani, A., A. Roshani, A. Roshani, M. Salehi, and A. Esfandyari. 2013. “A Simulated Annealing Algorithm for Multi-
manned Assembly Line Balancing Problem.” Journal of Manufacturing Systems 32 (1): 238–247.
Sawyer, J. F. H. 1970. Line Balancing. Washington, DC: Machinery and Allied Products Institute.
Scholl, A., M. Fliedner, and N. Boysen. 2010. “ABSALOM: Balancing Assembly Lines with Assignment Restrictions.”
European Journal of Operational Research 200: 688–701.
Sivasankaran, P., and P. Shahabudeen. 2014. “Literature Review of Assembly Line Balancing Problems.” International
Journal of Advanced Manufacturing Technology 73 (9–12): 1665–1694.
Tonge, F. M. 1961. A Heuristic Program of Assembly Line Balancing. Englewood Cliffs, NJ: Prentice-Hall.
Yilmaz, H., and M. Yilmaz. 2015. “Multi-manned Assembly Line Balancing Problem with Balanced Load Density.”
Assembly Automation 35 (1): 137–142.
Yilmaz, H., and M. Yilmaz. 2016. “A Multi-manned Assembly Line Balancing Problem with Classified Teams: A New
Approach.” Assembly Automation 36 (1): 51–59.
Zamzam, N., Y. Sadek, N. Afia, and A. El-Kharbotly. 2015. “Multi-manned Assembly Line Balancing Using Genetic
Algorithm.” International Journal of Engineering Research & Technology 73 (9–12): 1665–1694.