Professional Documents
Culture Documents
AltaraziYasin IPPSSDSTElectricalIndustry
AltaraziYasin IPPSSDSTElectricalIndustry
AltaraziYasin IPPSSDSTElectricalIndustry
net/publication/283094017
CITATION READS
1 14,289
2 authors:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Safwan A. Altarazi on 25 December 2015.
s_altarazi@asu.edu.jo
2
German-Jordanian University, Industrial Engineering Department, Amman, Jordan
{safwan.altarazi, o.yasin}@gju.edu.jo
Abstract. This paper addresses the integration of process planning and schedul-
ing (IPPS) with sequence dependent setup times for a case from the electrical
wires and power cables industry. While the IPPS problem has been a subject of
researchers’ attention in recent years, majority of research in this field neglect
setup time or assume it as a part of processing time. The objective is to simulta-
neously select the most feasible process plan and schedule with minimum
makespan. The problem is modeled as a mixed integer linear programming
problem and an example is presented to demonstrate the applicability and effec-
tiveness of the proposed modeling approach.
Different versions and characteristics of the IPPS problem associated with various
perspectives and solution techniques have been reported in literature. Kumar and
Rajotia [12] proposed a framework for the IPPS problem in a job shop environment
for axisymmetric components. Design specifications, availability of machine tools,
and shop floor status were considered in generating feasible process plans for incom-
ing products. Tan and Khoshnevis [23] proposed a linearized polynomial mixed inte-
ger model which assures that no overlapping in scheduling machines of operations
from different process. Garcia-Sabater et al. [6] presented a two-stage sequential
planning scheme for integrated operations planning and scheduling system using a
mixed-integer programming model. Altarazi [2] presented proposed a novel modeling
for the IPPS problem in job shop environment that simultaneously allocates opera-
tional tolerances while minimizing its manufacturing cost, minimizing work in pro-
cess inventory, and figuring operation-machine assignments. The proposed mixed-
integer nonlinear model minimized two objectives: the total operational tolerance-
manufacturing cost and the WIP inventory. A preemptive method was used to opti-
mize the two associated objectives. Baykasoğlu and Özbakır [4] also addressed the
IPPS problem with two objectives: total flow time and total cost of process plans. In
the proposed model, the generic process plan was represented as a grammar, dispatch-
ing rules were applied for sequencing operations, and a multi-objective Tabu search
was employed to generate alternative solutions. Li and McMahon [14] utilized simu-
lated annealing (SA) to facilitate the integration and optimization of process planning
and scheduling. The balanced level of machine utilization, makespan, job tardiness
and manufacturing cost were used to evaluate the approach. Brandimarte and
Calderini [5] implemented a two-phase hierarchical Tabu search for the integration
and optimization of the process planning and scheduling. Shao et al. [20] presented a
modified Genetic Algorithm (GA) approach for the IPPS problem where efficient
genetic representations and operator schemes have been developed. The experimental
results indicated the superiority and adaptability of the method. In addition, Amin-
Naseri and Afshari [3] presented a GA-based algorithm for the IPPS problem with
precedence constraints. Lee and Kim [13] introduced a simulation based GA approach
for solving the IPPS problem. In the presented approach, a simulation module com-
putes performance measures based on process plan combinations, then, these
measures are fed into a GA in order to improve the solution quality represented by
scheduling objectives such as makespan or lateness. Haddadzade, Razfar, and Zarandi
[8] considered stochastic processing times in their modeling of the IPPS problem and
solved it with a hybrid SA-Tabu algorithm. Guo et al. [7] used the particle swarm
optimization algorithm to solve the IPPS problem. Wong et al. [26] considered the
IPPS problem with rescheduling and used a hybrid multi-agent approach to solve this
problem in job shop environments. Kim et al. [11] utilized symbiotic evolutionary
algorithm, an artificial intelligence search technique, to handle the process planning
and scheduling functions simultaneously.
The majority of IPPS research neglect setup time consideration and or assume it as
a part of processing time. Recently, Wan et al. [25] implemented an Ant Colony algo-
rithm to solve the IPPS-SDST problem. Also, Nourali et al. [17] presented a mathe-
matical model to solve the IPPS-SDST problem. This research presents a new math-
ematical modeling for the IPPS-SDST considering a case from the electrical wires
and power cable industry.
The next section describes the production process of electrical wires and power ca-
bles. Section three is the statement of the problem while section four presents the
mathematical modeling for the problem. Implementation is presented and discussed in
section five. Finally, concluding remarks are presented in section six.
The electrical wires and cables industry is a crucial industry with hundreds of billions
investment around the world. Its’ products resemble the raw material for many other
industries business activities such as construction, telecommunication, power trans-
mission, electronics, etc. The products of the industry are usually classified into elec-
trical wires and cables. An electrical wire is a single flexible strand or solid rod of
metal, usually cylindrical, and surrounded by an insulator. Electrical cables on the
other hand, frequently called power cables, consist of two or more wires running next
to each other and bonded, braided, or twisted together, and commonly insulated to
form a single assembly.
An extruded cable production line is a sophisticated manufacturing process. It con-
sists of many sub processes that must work in concert with each other. The conductor
rod, usually from copper or aluminum, is first drawn to the specified diameter. After
drawing, the wire is softened, or annealed in a water bath. To raise the temperature for
the annealing process, a large electrical current passes through the wire for a fraction
of a second, raising its temperature briefly to about 1000F. Next, the wire, now soft
and flexible, is passed through an extruder, where either a single or double coating of
plastic insulation is applied. High-density polyethylene or polyvinyl chloride (PVC)
polymers, colored in one of ten industry-standard colors, are two typical polymers
used for this purpose. Exiting the extruder, the coated wire travels and passes through
another cooling trough and is coiled on take-up reels. The manufacturing steps for
single-rod wire ends by this stage. For two-rod wire and cables manufacturing two or
more conductors of the same gauge are twisted together forming what is called the
core. Next, a metal or non-metal mesh is braided around the cable. If the unit is to
form part of a larger cable, it next goes to the cabling operation. At cabling, multi
cores are twisted together on a rotating rod to form a multi-unit cable core. Depending
on the cable design and application, a protective metal sheathing of either aluminum
or aluminum and steel combined may be added in a manufacturing step called jacket-
ing or armor. Then, the outer cable jacket is insulated by black-color low-density
polyethylene, or cross-linked polyethylene (XLPE), through an extrusion process.
Finally, the jacketed cable then passes through a temperature- controlled water trough,
which cools the jacket. The cable is dried, and the top layer of the jacket is heated
slightly so that printer markings can be imprinted on it. Figure 2 shows the production
process flow of different wire and cable types.
Typically, the manufacturer of electrical wires and power cables faces a scenario
where multiple products, some with more than one feasible process plan, are to be
allocated on limited number of available machines and be produced within some time
limit. Furthermore, different allocation sequences of products (wires and cables) on
machines require a lot of setup efforts in terms of die changes in the drawing opera-
tion, extruders’ cleaning and changes of insulation material or color for the insulation
operation, parameters changes of the twisting and annealing operations, and other
setup time-consuming activities. The above scenario in fact perfectly fits within the
IPPS-SDST scope. The IPPS-SDST considers a set of n number of jobs J = {Jj}1≤ j≤ n
and a set of m number of machines M = {Mi}1≤i≤m. Each job Jj is defined as a set of
operations that have to be processed according to a set of r number of process plans L
= {Ll}1≤ l≤ r. The processing of job Jj on machine Mi in process plan Ll is called opera-
tion Ojil. Operation Ojil requires the exclusive use of Mi for deterministic processing
time pji , i.e., each machine can process, at most, one operation at a time. This prob-
lem is strongly NP-hard since its simpler version with 0 setup times and 1 process
plan per job is also NP-hard.
Fig. 2. The manufacturing process flowcharts for examples of electrical wires and power cable
[22]
4 Mathematical Modeling
4.1 Notation
N total number of jobs
M total number of machines
R total number of process plans
nj total number of operations of job j
Oji ith operation of job j
Ojil ith operation of job j in process plan l
pji processing time of Oji
Si the set of operations assigned to machine Mi
xji the starting time of the processing of operation Oji
xjil the starting time of the processing of operation Ojil on process plan l
Cmax makespan of a schedule
K An arbitrary large positive number
Sjki setup time between job j and job k if job j precedes job k on machine i
1, if operation 𝑂𝑘𝑖 precedes operation 𝑂𝑗𝑖 on machine 𝑖
𝑦𝑗𝑘𝑖 = {
0, otherwise
𝑦𝑗𝑘𝑖𝑙𝑓
1, if operation 𝑂𝑘𝑖𝑓 on process plan 𝑓 precedes operation 𝑂𝑗𝑖𝑙 on process plan 𝑙
= {
0, otherwise
1, if process plan 𝑙 is selected for job 𝑗
𝑑𝑗𝑙 = {
0, otherwise
Note that the xji and xjil are the continuous decision variables and 𝑦𝑗𝑘𝑖 , 𝑦𝑗𝑘𝑖𝑙𝑓 and
𝑑𝑗𝑙 are the binary decision variable.
∑ 𝑑𝑗𝑙 = 1 ∀ 𝑗 = 1, … , 𝑛 (3)
𝑙=1
𝑥𝑗𝑖𝑙 − 𝑥𝑘𝑖𝑓 + 𝐾(1 − 𝑦𝑗𝑘𝑖𝑙𝑓 ) + ∀ (𝑘, 𝑖), (𝑗, 𝑖) ∈ 𝑆𝑖 , (𝑘, 𝑖) ≠ (𝑗, 𝑖), 𝑖 =
𝐾(1 − 𝑑𝑗𝑙 ) ≥ 𝑝𝑘𝑖𝑓 + 𝑆𝑘𝑗𝑖 1, … , 𝑚; 𝑙, 𝑓 = 1, … , 𝑟 (5)
𝑥𝑗𝑖𝑙 ≥ 𝑆0𝑗𝑖 ∀ 𝑗 = 1, … , 𝑛; 𝑖 = 1, … , 𝑛𝑗 ; 𝑙 = 1, … , 𝑟
(7)
𝑥𝑗𝑖 ≥ 0 ∀ 𝑗 = 1, … , 𝑛; 𝑖 = 1, … , 𝑛𝑗
(8)
The objective function (1) is the minimization of makespan Cmax. Constraints set (2)
ensures that jobs are processed according to precedence relationship between opera-
tions of the same job, stating that an operation Oj,i+1,l cannot start before the end of its
preceding operation Ojil in job Jj. Constraints set (3) ensures that only one process
plan is selected for job j. Constraints sets (4) and (5) are called disjunctive constraints
because one or the other alone must hold for the selected process plan only, i.e., the
two constraints hold only for one process plan per job and, each two distinct opera-
tions Ojil and Okif sharing the same machine i on in the selected process plan cannot be
scheduled simultaneously. Also, these two constraints are used to incorporate the
sequence dependent setup times. Constraints set (6) defines the maximum completion
time of all jobs, i.e., the makespan of the schedule. Constraints set (7) enforces the
start time of any operation to occur after the initial setup time on its first assigned
machine ,i.e., setup time for each job j on when it comes first on the sequence of jobs
on machine i. Constraints sets (8-10) imply non-negativity and integrality of the cor-
responding variables.
5 Example
Table 2. Processing times for the four products on the four machines.
Processing Time
Job/machine 1 2 3 4
1 2 3 2 3
2 3 2 7 2
3 4 3 6 4
4 10 3 4 5
Table 3. Setup times for the four products on the four machines
Machine 1 Machine 2
Job Job
► 1 2 3 4 ► 1 2 3 4
▼ ▼
0 1 1 1 1 0 2 2 2 2
1 0 1 2 0 1 0 0 1 1
2 1 0 1 0 2 0 0 1 0
3 1 0 0 1 3 0 2 0 0
4 1 0 2 0 4 0 2 1 0
Machine 3 Machine 4
Job Job
► 1 2 3 4 ► 1 2 3 4
▼ ▼
0 3 3 3 3 0 4 4 4 4
1 0 0 2 1 1 0 3 1 3
2 1 0 1 1 2 2 0 2 2
3 0 2 0 1 3 1 4 0 3
4 0 2 0 0 4 1 1 2 0
Using the inputs shown in Tables 1-3, the MIP model shown in equations (1-10) was
developed under AMPL IDE and solved using CPLEX 12.6.0.0 solver running on a
PC with a core i5 2.27 GHz and 4GB RAM. The AMPL code is given by the appen-
dix. The resulted optimal solution, represented by the starting time of each job on
each machine along with the selected process plan for each job, is given by Table 4.
As can be seen, the first process plan was selected for jobs 1, 2 and 4, and the second
process plan was selected for job 3. A Gantt chart for the optimal schedule, including
processing and setup times for each job according to the selected process plans, is also
shown in figure 3. As can be seen, the makespan is 27 time units.
For comparisons purposes, the above case was solved as a scheduling problem on-
ly in which only single process plan (operations sequence) for each job was assumed
available. After running the model, the makespan was found to be 28 and 34 time
units when considering separately the first and the second process plans, respectively.
This indeed approves the benefits of integrating the IPPS and the SDST problems and
shows the superiority which the IPPS-SDST modeling can provide over separate im-
plementation of the SDST problem.
Starting Time
Operation Selected process plan
Job 1 2 3 4
1 19 14 11 8 First
2 14 17 20 4 First
3 23 20 3 12 Second
4 1 11 14 22 First
6 Conclusions
References
1. Allahverdi, A., Ng, C.T., Cheng, T.E., Kovalyov, M.Y.: A survey of scheduling problems
with setup times or costs. Europ J Oper Res, 187, 985-1032 (2008)
2. Altarazi, S.: An optimization tool for operational tolerances allocation, work in process in-
ventory minimization, and machines assignment in a discrete part manufacturing environ-
ment. Int J Adv Manuf Technol, 55, 1069-1078 (2011)
3. Amin-Naseri, M.R., Afshari, A.J.: A hybrid genetic algorithm for integrated process plan-
ning and scheduling problem with precedence constraints. Int J Adv Manuf Technol, 59,
273-287 (2012)
4. Baykasoğlu, A., Özbakır, L.: A grammatical optimization approach for integrated process
planning and scheduling. J Intell Manuf, 20, 211-221 (2009)
5. Brandimarte P, Calderini M (1995) A hierarchical bicriterion approach to integrate process
plan selection and job shop scheduling. Int J Prod Res 33(1):161–181
6. Garcia-Sabater, J.P., Maheut, J., Garcia-Sabater, J.J.: A two-stage sequential planning
scheme for integrated operations planning and scheduling system using MILP: the case of
an engine assembler. Flex Serv Manuf J, 24, 171-209 (2012)
7. Guo, Y.W., Li, W.D., Mileham, A.R., Owen, G.W.: Applications of particle swarm opti-
mization in integrated process planning and scheduling. Robot Comput-Integr Manuf, 25,
280-288 (2009)
8. Haddadzade, M., Razfar, M.R., & Zarandi, M.F.: Integration of process planning and job
shop scheduling with stochastic processing time. Int J Adv Manuf Technol, 71, 241-252
(2014)
9. Jain, A., Jain, P.K., & Singh, I.P.: An integrated scheme for process planning and schedul-
ing in FMS. Int J Adv Manuf Technol, 30, 1111-1118 (2006)
10. Kang, M., Han, J., Moon, J.G.: An approach for interlinking design and process plan-
ning. J. Mater. Process. Technol, 139, 589-595 (2003)
11. Kim, Y. K., Park, K., Ko, J.: A symbiotic evolutionary algorithm for the integration of
process planning and job shop scheduling. Comput Oper Res, 30, 1151-1171 (2003)
12. Kumar, M., Rajotia, S.: Integration of process planning and scheduling in a job shop envi-
ronment. Int J Adv Manuf Technol, 28, 109-116 (2006)
13. Lee H, Kim SS (2001) Integration of process planning and scheduling using simulation
based genetic algorithms. Int J Adv Manuf Technol (2001) 18:586-590
14. Li, W.D., McMahon, C.A.: A simulated annealing-based optimization approach for inte-
grated process planning and scheduling. Int J Comput Inetll Manuf, 20, 80-95 (2007)
15. Li, X., Gao, L., Shao, X., Zhang, C., Wang, C.: Mathematical modeling and evolutionary
algorithm-based approach for integrated process planning and scheduling. Comput Oper
Res, 37, 656-667 (2010)
16. Mandahawi, N., Al-Shihabi, S., Altarazi, S.: A max-min ant system to minimize total tar-
diness on a single machine with sequence dependent setup times implementing a limited
budget local search. Int J Res Rev App Sci, 6, 30-40 (2011)
17. Nourali, S., Imanipour, N., Shahriari, M. R.: A Mathematical Model for Integrated Process
Planning and Scheduling in Flexible Assembly Job Shop Environment with Sequence De-
pendent Setup Times. Int J Math Analy, 6, 2117-2132 (2012)
18. Pinedo, M.L.: Scheduling: theory, algorithms, and systems. Springer Science & Business
Media, NewYork (2012)
19. Saygin, C., Kilic, S.E.: Integrating flexible process plans with scheduling in flexible manu-
facturing systems. Int J Adv Manuf Technol, 15, 268-280 (1999)
20. Shao X, Li X, Gao L, Zhang C (2009) Integration of process planning and scheduling-A
modified genetic algorithm-based approach. Comput Oper Res 36:2082-2096
21. Stafford, E.F., Tseng, F.T., Gupta, J.N.D.: Comparative evaluation of MILP flowshop
models. J Oper Res Soc, 56, 88-101 (2005)
22. Taiwan Turnkey Project Association,
http://www.tpcc.org.tw/wholeplant2/mainchoose/1-4.htm
23. Tan, W., Khoshnevis, B: A linearized polynomial mixed integer programming model for
the integration of process planning and scheduling. J Intell Manuf, 15, 593-605 (2004)
24. Tan, W., Khoshnevis, B.: Integration of process planning and scheduling—a review. J In-
tell Manuf, 11, 51-63 (2000)
25. Wan, S., Wong, T., Zhang, S., Zhang, L.: Integrated process planning and scheduling with
setup time consideration by ant colony optimization: Doctoral dissertation. The University
of Hong Kong, Hong Kong (2012)
26. Wong, T.N., Leung, C.W., Mak, K.L., Fung, R.Y.K.: Integrated process planning and
scheduling/rescheduling—an agent-based approach. Int J Prod Res, 44, 3627-3655 (2006)