See discussions, stats, and author profiles for this publication at: [Link]

net/publication/318689092

Simulation modelling and analysis of a production line

Article in International Journal of Simulation and Process Modelling · January 2017

DOI: 10.1504/IJSPM.2017.10006536

CITATIONS READS

13 993

3 authors, including:

Mahmoud El-Sharief Mahmoud Heshmat

Assiut University Assiut University
36 PUBLICATIONS 214 CITATIONS 50 PUBLICATIONS 302 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Mahmoud El-Sharief on 10 May 2023.

The user has requested enhancement of the downloaded file.

Int. J. Simulation and Process Modelling, Vol. 12, Nos. 3/4, 2017 369

Simulation modelling and analysis of a production

line

Mahmoud Heshmat*, Mahmoud El-Sharief and

Mohamed El-Sebaie
Mechanical Engineering Department,
Assiut University, Egypt
Email: mheshmat@[Link]
Email: msharif@[Link]
Email: mgad@[Link]
*Corresponding author

Abstract: Production lines modelling has many problems that are difficult to be solved using
analytical solutions, due to uncertainty and/or variability in variables and parameters. This paper
is targeting unreliable production lines with finite buffers for the objective of evaluating and
analysing the current situation and identifying bottlenecks. We use discrete event simulation to
model a real production line based on one-year historical data about the breakdowns of all the
production line’s machinery. This data is used to find appropriate probability distributions to
represent each machinery downtime and uptime. The simulation model is built using AnyLogic
software to abstract the actual production line, and the model is validated using the actual data
and information about the production line. Improvement scenarios are proposed to resolve the
observed bottlenecks, and therefore give managerial insights to increase the throughput according
to the increasing demand. Finally, production plans are set for different demands along the year.

Keywords: production lines; simulation; cement production; AnyLogic.

Reference to this paper should be made as follows: Heshmat, M., El-Sharief, M.A. and
El-Sebaie, M.G. (2017) ‘Simulation modelling and analysis of a production line’,
Int. J. Simulation and Process Modelling, Vol. 12, Nos. 3/4, pp.369–376.

Biographical notes: Mahmoud Heshmat is a PhD student at the Department of Industrial

Engineering and Systems Management in Egypt-Japan University of Science and Technology
(E-JUST), Alexandria, Egypt. Currently, he is on leave from Assiut University. He earned his
BSc and MSc in Mechanical Engineering, Design and Production Section from Assiut
University, Egypt. His research interests include operations research, simulation, optimisation,
healthcare management, scheduling, and manufacturing.

Mahmoud El-Sharief is an Associate Professor at the Mechanical Engineering Department at

Assiut University, Assiut, Egypt. His research interests include operations research, simulation,
project management, and mechanical engineering.

Mohamed El-Sebaie is an Emeritus Professor at the Mechanical Engineering Department at

Assiut University, Assiut, Egypt. His research interests include metal forming, machine design,
operations research, and mechanical engineering.

1 Introduction any current practice by measuring the key performance

indicators (KPIs) such as throughput, buffer capacities,
Production lines have been treated densely in the literature
resources use, work in process, and waiting times.
using mathematical modelling and simulation modelling.
Discrete event simulation is used effectively to capture
However, most of the developed mathematical models have
and analyse complex manufacturing environment and
unrealistic assumptions, such as reliable workstations,
interactions with other complex systems. This can be done
infinite buffers, synchronous line, and identical
with less effort compared to analytical models. It can be
workstations. On the other hand, simulation is applied to
efficiently used to make perfect and thorough decisions to
model and analyse real-world production lines in a flexible
identify the best alternative among several candidates
way by conducting experiments and testing different
(Aqlan, Lam, & Ramakrishnan, 2014). In their literature
production scenarios. Simulation models also can evaluate
review, Hosseinpour et al. (2009) confirmed the importance
existing production lines, and explore bottlenecks
of simulation in manufacturing as a helpful tool in the
effectively. Moreover, they can measure the performance of

Copyright © 2017 Inderscience Enterprises Ltd.

370 M. Heshmat et al.

industrial field. Simulation is low cost, secure and fast

analysis tool with many different system configurations. for each workstation. Furthermore, one-year historical data
In terms of production lines, simulation can be applied of failure and repair are filtered and analysed for each
to estimate the size of buffers and evaluate the effect of machinery, including preventive and predictive
adding a new piece of equipment on an existing production maintenance, annual shut down, workstation blocking and
line. Moreover, it can be used to analyse and estimate the starvation, and even workers strike. Then, it is used to
throughput, find, analyse, and resolve bottlenecks, estimate determine the best probability distribution considering
times which parts spend in queues, and queue size, and statistical evidence. The simulation model is built using
estimate utilisations of workstations. Papadopoulos et al. AnyLogic software (AnyLogic, 2016), in which each
(2009) stated that “Simulation of production lines is a workstation is represented with the actual processing time,
powerful tool in obtaining the performance measures where failure time, repair time, and time between failures
analytical methods are either difficult or impossible to use”. distribution.
Therefore, simulation is adopted in this paper to model and The remainder of the paper is organised as follows:
analyse a production line because of its relative advantages. Section 2 describes the whole production line from the first
Simulation is applied extensively in the literature to stage until the final product packaging stage. Section 3
model, analyse, and balance flow lines. For example, Kayar presents the simulation model in detail according to
and Akalin （2016） developed a discrete-event simulation methodology. Section 4 gives the results of the
simulation model in order to compare heuristic to simulation model, the observed bottleneck, and its solution.
simulation. Abidi et al. (2015) used simulation alongside Section 5 gives production plans according to different
virtual reality in lean manufacturing. Omar et al. (2016) demands.
used a hybrid simulation approach from discrete and
continuous perspectives in order to predict energy flow in
production lines. Jeon and Kim (2016) reviewed different 2 The production line description
simulation methods, namely discrete event simulation,
Figure 1 shows a schematic diagram of the cement
agent-based simulation, and system dynamics simulation in
production line with its workstations. First, trucks load
production planning, and concluded that discrete event
limestone and clay concurrently from the quarry area and
simulation is the most used tool in that area. Atieh et al.
transfer them to crushers. After crushing, the crushed
(2015) used simulation besides value stream mapping to
limestone and clay are transferred separately by belt
detect bottlenecks in an application in the glass industry.
conveyors to the stacking area and by two workstations
Fanti et al. (2015) integrated simulation Simulation
called stacker and reclaimer, the crushed limestone and clay
modelling and analysis of a production line with
are stockpiled in a scattered way to make them
mathematical models to solve scheduling problems in a
homogenous. After that, conveyor belts transfer the
steel making factory. Huang et al. (2016) integrated
limestone and clay independently to the raw mill area where
simulation with the neural network concept in order to
two raw mills exist. The two raw mills grind limestone and
predict the throughput of the production line. Sajadi and
clay together until they become a fine homogeneous
Rad (2016) used simulation optimisation to find the best
mixture. After the milling process, the output mixture is
production rate. Galankashi et al. (2016) used simulation
transferred to raw-mill silos before entering the kiln to
and design of experiments to evaluate the performance of a
produce what so called clinker. After forty minutes inside
petrol station. Guner et al. (2016) used simulation to
the kiln, the clinker is cooled inside a cooler. The cooled
anticipate maintenance decisions for production line
clinker is transferred and stocked in clinker stocks. Then,
machines. Pawlewski and Kluska (2014) used both agent-
the clinker is transferred from the stock and grinded using
based simulation and discrete-event simulation to simulate a
cement mills and then stocked in cement silos as a final
bus assembling line. Hsieh et al. (2016) used simulation to
product. Finally, the cement is discharged from the silos to
model a case of integrated circuit packaging. Pekel and
be packaged using packaging machines and eventually to
Kara (2016) used discrete event simulation to make bus
the customer trucks.
scheduling.
This cement production line works 24 hours a day and
In this paper, a cement production line is studied and
seven days per week. The crushers, stacker, and reclaimer
monitored for one year. Actual data is collected including
work two shifts per day (18 hours) and the remaining
production capacities, transfer times, and processing times
workstations work 24 hours per day.
 Simulation modelling and analysis of a production line 371

Figure 1 Schematic diagram of the cement production line (see online version for colours)

Source: Heshmat et al. (2013)

3 The simulation model 3.2 Workstations downtime and uptime distributions

This section presents the simulation model, first, the The recorded data gives a detailed schedule for each
simulation methodology is given, then, the downtime and workstation, identifying the downtime duration along one
uptime distributions for each machinery are fitted, after that year. The data was collected according to a spreadsheet
the model verification and validation are performed. template contains every component in the production line.
Every area operator recorded any breakdown event occurred
3.1 The simulation methodology to each component. For instance, Table 1 gives a sample of
the breakdown occurrences for one of the two raw mills for
We follow a clear methodology to conduct the simulation one month. It can be seen that every failure is recorded with
experiments as shown in Figure 2. First, the production line its duration. All the stoppages occurrences for the remaining
is studied in detail in the field and by collecting data about equipment are collected with the same way.
each workstation including its processing time, historical The downtime data is deeply detailed since it contains
records for one year of failure, repair, and time between times of preventive and corrective maintenance, shutdown,
successive failures. The historical data is fitted using failures, repair, and strikes. ARENA Input Analyser is used
ARENA input analyser to find the best probability for choosing the best probability distribution for the
distribution of downtime and uptime for each workstation. breakdowns of each workstation. The resulted probability
In order to well-fit the obtained probability distribution, we distributions are tested with the two standard measures of
revise the data to remove outliers and collect more data if distribution’s fitting, namely the Chi-square test and
needed, then the best probability distributions are obtained Kolmogorov-Smirnov (K-S) goodness-of-fit hypothesis test.
with high statistical confidence. Second, we start building Table 2 summarises the best-fitted probability distribution
the simulation model using AnyLogic software. During this of the downtime for each workstation with the
stage, we contacted with the cement production line corresponding p-value. Regarding the packaging machines,
engineers to get any missing and needed data. Third, we there is a standby machine for each one, so the downtime
make a pilot run to ensure that the logic of the model is distributions do not be considered since the packaging
satisfied. Fourth, we verify the model using face validity process continues without stopping.
which depends on the model animation. Fifth, we validate In terms of the uptime distributions, the time between
the model using walkthrough by comparing the model breakdowns is calculated for the recorded breakdown data
throughput with the actual throughput and checking the using Microsoft excel, and then ARENA Input Analyser is
significant difference between the two means using the t- used to fit the times between breakdowns for every
test. Once the model is valid, we investigate bottlenecks workstation. Table 3 summarises the uptime distributions
observed in the production line and doing the needed for the production line equipment with the corresponding p-
experiments for improving the KPIs and setting any values. Simulation modelling and analysis of a production
production plan. Simulation modelling and analysis of a line.
production line.
372 M. Heshmat et al.

Figure 2 A flowchart represents the simulation methodology Table 1 A sample of the collected data for one of the two raw
(see online version for colours) mills

Downtime
Day Stoppage type Cause
(hours)
01 Energy saving1 4.75 Energy saving during peak
time
01 Electrical 0.75 Electric problem in the
problem clay reclaimer (The touch
screen was broken)
01 Electrical 0.5 Too long start of rheostat
problem of recycle
02 Electrical 1.83 Too long start on the
problem rheostat of mill fan
02 Production 0.5 Rope switch of main
problem feeding belt conveyor
02 Mechanical 0.75 Tighten bolts in the mill
problem shell
03 Electrical 3.167 Displacement of clay
problem reclaimer
03 Energy saving 4.25 Energy saving during peak
time
04 Electrical 2.25 Trip mill fan
problem
05 Lack of own 2.25 End limestone pile and
process changing to new pile
materials
08 Mechanical 8.5 Leakage of oil from outlet
problem low-pressure pump
09 Mechanical 2.5 Leakage of oil from outlet
problem low-pressure pump from its
oil seal
11 Energy saving 4.25 Saving power at peak time
18 Mechanical 7.167 Vulcanising clay belt
problem conveyor
18 Mechanical 3.83 Clay belt conveyor
problem
18 Energy saving 3.3 Energy saving during peak
time
19 Mechanical 2 Mechanical problem
problem in the chain of clay belt
conveyor
20 Production 1 Blockage in weigh feeder
problem
20 Mechanical 1 As a request of mechanical
problem team to check clay
reclaimer
26 Mechanical 0.67 High-temperature inlet
problem bearing
Note: 1At peak times, operations managers shut down
some equipment to save electrical energy.
 Simulation modelling and analysis of a production line 373

Table 2 The downtime distributions of the production line’s workstations

Workstation Fitted probability distribution (hrs) p-value

Limestone crusher Gamma (12.3, 0.69) > 0.15
Clay crusher 0.999 + gamma (12.7, 0.633) > 0.15
Limestone stacker Weibull (2.82, 0.598) > 0.15
Clay stacker Weibull (4.43, 0.705) > 0.15
Limestone reclaimer Weibull (4.03, 0.687) > 0.15
Clay reclaimer 0.001 + Weibull (4.86, 0.687) > 0.15
Raw mill 1 Weibull (8.64, 1.27) > 0.15
Raw mill 2 Weibull (7.95, 1.19) > 0.15
Kiln Uniform (10, 24) 0.13
Cooler Uniform (10, 24) 0.56
Cement mill 1 Lognormal (4.85, 10.4) 0.13
Cement mill 2 Lognormal (4.69, 10.1) > 0.15
Cement mill 3 Lognormal (5.1, 9.4) 0.1

Table 3 The uptime distributions of the production line’s workstations

Workstation Fitted probability distribution (hrs) p-value

Limestone crusher 9 + 13 * beta (1.68, 1.36) > 0.75
Clay crusher 9 + 13 * beta (1.69, 1.4) > 0.15
Limestone stacker 10 + 12 * beta (1.47, 1.25) > 0.15
Clay stacker 10 + 12 * beta (1.12, 1.06) 0.374
Limestone reclaimer 10 + 12 * beta (1.41, 1.3) > 0.15
Clay reclaimer 10 + 12 * beta (1.25, 1.11) > 0.15
Raw mill 1 48 * beta (0.99, 2.69) > 0.15
Raw mill 2 Weibull (7.95, 1.19) > 0.15
Kiln 21 + exponential (200) > 0.15
Cooler 21 + exponential (202) > 0.15
Cement mill 1 6 + lognormal (122, 466) > 0.15
Cement mill 2 22 + exponential (123) 0.14
Cement mill 3 20 + exponential (119) 0.12

3.3 The number of replications

Figure 3 Mean and standard deviation of the throughput versus
The Welch rules are applied to determine the minimum number of replication (see online version for colours)
number of replications for each simulation experiment. We
120
choose the throughput as a testing parameter to find the Mean
average number of replications versus the replication 100 Standard Deviation
Throughput (per 50 tons)

number. Figure 3 depicts both the average and the standard

80
deviation of the throughput versus the number of
replications. It is clear from the figure that the variability of 60
the average throughput and the standard deviation begins to
40
be steady after the seventh replication.
The considered production line is steady state since it 20

works continuously for 24 hours a day for seven days a

0
week. The simulation model is run for six months with ten 0 1 2 3 4 5 6 7 8 9 10
replications and the warm-up period is set as 24 hours. Number of Replications

Figure 4 depicts a snapshot of the simulation model which

is built in AnyLogic 7.3.6. The model mimics the standing
line from limestone and clay arrivals in the left until the
packing process and then the entities go to the sink.
374 M. Heshmat et al.

Figure 4 A snapshot of the developed simulation model (see online version for colours)

Table 4 The results of the t-test

Figure 5 A snapshot of the t-test results using Minitab 17
Actual throughput Throughput from simulation
(tonne/day) (tonne/day)
4,710 4,295
5,180 4,820.7
4,820 4,010.5
5,170 4,845
4,004 4,200.1
4,090 4,710 4 Results and discussion
5,230 4,345.9
This section aims at presenting and discussing the
4,796 4,007.2 simulation results. Investigation of the production line
machines utilisations reveals that the kiln, cooler, and raw
3.4 The model validation mills have the highest utilisation as shown in Figure 6. This
means that these areas are congested with work in process
As validation is a crucial step in simulation modelling, first, which increases the chance of blockage. Regarding this
the animation method is used to ensure that the model is run point, we have been given a feedback from the production
without errors. All the entities, resources, flow, and line’s operations engineers that the kiln area always has
throughput imitate the actual production line from raw blocking problems.
material arrival until the final product packaging stage, in
logic, precedence, and movement, therefore the model is Figure 6 Utilisations of the production line’s machinery
verified. After that, we validate the model by comparing its (see online version for colours)
outputs with the real production line outputs. The average 100
daily throughput of the model is compared with its peer in 90
80
the actual production line, that is the number of tonnes
Utilization %

70
60
produced per day. The average daily throughput of the 50
40
simulation model is 4,404.3 tonnes per day, whereas, the 30
20
average throughput of the real production line is about 10
0
4,750 tonnes per day without a significant difference
s
i ln
ler

er

ne
er

er

r
er
s

ill

between the two means according to the t-test results.

ke

ke
i ll

im

sh
sh

hi
K

m
oo

M
M

ac

ac
ru

ac
i
ru
la

la
C

aw
St

St
t

ec
C

M
C

ec
en

Table 4 gives both the actual and simulated throughputs

R
y

ne
R
y

R
ne
em

g
la
la

in
y

to
ne
C

to
C

la
C

ck
es
es

to
C

Pa

which we apply the t-test upon. The result of the t-test is

m

es

Li
Li

m
Li

demonstrated in Figure 5 which ensures that there is no

significance difference between the two values, namely the The second key factor is the buffer size. Most of the buffer
actual and simulated throughputs. This is clear from the p- sizes are observed to be below the maximum limit except
value (0.1), it is greater than 0.05. Therefore, the simulation for the raw mills silos, which has overstock problems
model is valid. observed in the simulation runs. This overstock problem is
 Simulation modelling and analysis of a production line 375

because the kiln cannot accommodate all the arriving raw and the buffer size before the kiln will decrease as well. If
material due to its limited capacity. Another KPI, we run the any starvation occurs before the kiln area, then the crusher,
simulation model without considering breakdowns of the stacker, and reclaimer can be operated 24 hours other than
workstations to know the theoretical throughput of the 18 hours/day. In that case, simulation modelling can also be
production line, and therefore, the production line efficiency adopted to analyse that point.
can be estimated. The throughput without considering
breakdowns is 5,120 tonnes per day whereas; it is 4,404.3
tonnes per day in case of breakdown consideration. Hence, 5 Production planning for different demands
the production line efficiency is 0.86.
So far the executive board could put production planning
scenarios to face the fluctuating demand. This section is
4.1 Observed bottlenecks dedicated to using simulation to set the production line to
We use the definition of a bottleneck from Yu and Matta accommodate different demands along the year. The large
(2016) which states that “a machine becomes the bottleneck space in the actual production line is exploited to determine
if the sensitivity of the system throughput to its throughput any needed inventory. Let the expected demand is as listed
in isolation is higher than all other machines in the system”. in Table 5, and these targeted demands require increasing
It is clear from the simulation outputs and this rule that there the capacity of some workstations. Several simulation
is a bottleneck in the clinker production area. That is experiments are conducted to reach to the production line
because of the highest utilisation of the kiln (95%) and also capacity modifications.
the largest queue before the kiln. This big queue makes
Table 5 Demands and corresponding modifications in the
blockages before the kiln and therefore raises the average
production line
utilisation of the kiln. In addition, the operations managers
have informed us that they have to shut down the raw mills Demand (tonne/day) Modifications
for hours due to the high silo level. This feedback increases 500 No
the confidence in our simulation model results. On the other
600 Increasing the kiln capacity by 20%
hand, quite low utilisation in the succeeding stages, namely
and operating the crusher, stacker, and
the cement mills and the packaging machines. We have reclaimer for 20 hours instead of 18
been also informed by the operations engineers that hours per day
sometimes there is starvation in the cement mills which 700 Increasing the kiln capacity by 30 %
enforce them to shut down one of them temporarily. This and operating the crusher, stacker, and
bottleneck also has an effect on the preceding buffer size reclaimer for 24 hours instead of 18
and makes blocking at all the preceding stages. This hours per day
problem enforces the operations managers to operate the 300 Working all the line for 18 hours
workstations before the raw mills only two shifts (18 hours) instead of 24 hours per day
instead of three shifts (24 hours). This action is considered
an indirect economic loss in these workstations due to the
idleness rather than adding value. Simulation modelling and 6 Conclusions
analysis of a production line.
The objective of this study was achieved by using discrete
event simulation to capture, analyse, and measure some
4.2 Bottleneck resolving
KPIs for a cement production line. The production line was
To resolve the observed bottleneck in the kiln area, a thoroughly analysed and found to have bottlenecks that
proposal is made to increase the kiln capacity by 10%. Then were causing congestion in the kiln area. Simulation was
the model is run with the new proposal and thus the used to analyse this bottleneck and resolve it. The
bottleneck is resolved, hence the total production line production line throughput could be increased by 8% if the
throughput increases by 8%. This improvement could be kiln capacity is increased by 10%. In addition, the
implemented in the actual production line because there is a simulation model can be used to check production scenarios
sufficient space after the kiln area apart from the clinker according to demand. The feedback obtained from the
stocks. Another alternative, the kiln could be replaced by production line operations managers supports the credibility
another larger one which has a larger diameter with the of the simulation model.
same length. The cost of the new extension or even the new An important extension to this work, an economic study
kiln could be compensated after a certain time especially the will be studied to find the cost of commissioning an
demand for the product has been increasing. Adding a new extension to the kiln area. This cost will be compared to the
extension of the kiln or replacing it will decrease the profit coming from increasing the throughput. That feasible
idleness of the succeeding states, and thus the throughput, study is considered a crucial step towards the
implementation phase.
376 M. Heshmat et al.

References Hsieh, L.Y., Huang, E., Chen, C-H., Zhang, S. and Chang, K-H.
(2016) ‘Application of multi-fidelity simulation modelling to
Abidi, M., Lyonnet, B., Chevaillier, P. and Toscano, R. (2015) integrated circuit packaging’, International Journal of
‘Contribution of virtual reality for lines production’s Simulation and Process Modelling, Vol. 11, Nos. 3/4,
simulation in a lean manufacturing environment’, pp.259–269.
International Journal of Computer Theory and Engineering,
Huang, P.B., Chen, J.C., Yu, C.W. and Chen, C.C. (2016)
Vol. 8, pp.182–189.
‘Integration of simulation and neural network in forecasting
AnyLogic (2016) ‘Multi-method simulation software’ [online] the throughput for TFT-LCD colour filter fabs’, International
[Link] (accessed 12 December 2016). Journal of Computer Integrated Manufacturing, Vol. 29, No.
Aqlan, F., Lam, S.S. and Ramakrishnan, S. (2014) ‘An integrated 3, pp.298–308.
simulation-optimization study for consolidating production Jeon, S.M. and Kim, G. (2016) ‘A survey of simulation modeling
lines in a configure-to-order production environment’, techniques in production planning and control (PPC)’,
International Journal of Production Economics, Vol. 148, Production Planning & Control, Vol. 27, No. 5, pp.360–377.
pp.51–61.
Kayar, M. and Akalin, M. (2016) ‘Comparing heuristic and
Atieh, A.M., Kaylani, H., Almuhtady, A. and Al-Tamimi, O. simulation methods applied to the apparel assembly line
(2015) ‘A value stream mapping and simulation hybrid balancing problem’, Fibres and Textiles in Eastern Europe,
approach: application to glass industry’, The International Vol. 24, No. 2, pp.131–137.
Journal of Advanced Manufacturing Technology, Vol. 84,
Omar, M.A., Qilun, Z., Lujia, F., Ali, A.A., Lahjouji, D. and
Nos. 5–8, pp.1573–1586.
Khraisheh, M. (2016) ‘A hybrid simulation approach for
Fanti, M.P., Rotunno, G., Stecco, G., Ukovich, W. and Mininel, S. predicting energy flows in production lines’, International
(2015) ‘An integrated system for production scheduling in Journal of Sustainable Engineering, Vol. 9, No. 1, pp.25–34.
steelmaking and casting plants’, IEEE Transactions on
Papadopoulos, C.T., O’Kelly, M.E.J., Vidalis, M.J., S. (2009)
Automation Science and Engineering, Vol. 13, No. 2,
Analysis and Design of Discrete Part Production Lines,
pp.1112–1128.
Springer, New York.
Galankashi, M.R., Fallahiarezoudar, E., Moazzami, A., Yusof,
Pawlewski, P. and Kluska, K. (2014) ‘Using DES/ABS approach
N.M. and Helmi, S.A. (2016) ‘Performance evaluation of a
to model and simulate bus assembling process’, Highlights of
petrol station queuing system: a simulation-based design of
Practical Applications of Scalable Multi-Agent Systems. The
experiments study’, Advances in Engineering Software, Vol.
PAAMS Collection: International Workshops of PAAMS
92, pp.15–26.
2016, Sevilla, Spain, 1–3 June 2016, Springer, Vol. 616,
Guner, H.U., Chinnam, R.B. and Murat, A. (2016) ‘Simulation pp.110–132.
platform for anticipative plant-level maintenance decision
Pekel, E. and Kara, S.S. (2016) ‘Simulation-based fleet scheduling
support system’, International Journal of Production
in the Metrobus’, International Journal of Simulation and
Research, Vol. 54, No. 6, pp.1785–1803.
Process Modelling, Vol. 11, Nos. 3/4, pp.326–336.
Hajihosseini, F.H. and Hajihosseini, H. (2009) ‘Importance of
Sajadi, S.M. and Rad, M.F. (2016) ‘Optimal production rate in
simulation in manufacturing’, World Academy of Science,
production planning problem with simulation optimisation
Engineering and Technology Vol. 51, pp.285–288.
approach by simulated annealing’, Int. J. Industrial and
Heshmat, M., El-Sharief, M.A. and El-Sebaie, M.G. (2013) Systems Engineering, Vol. 22, No. 3, pp.262–280.
‘Simulation modeling of production lines: a case study of
Yu, C. and Matta, A. (2016) ‘A statistical framework of data-
cement production line’, Journal of Engineering Sciences,
driven bottleneck identification in manufacturing systems’,
Vol. 41, No. 3, pp.1045–1053.
International Journal of Production Research, Vol. 7543,
No. 21, pp.1–16.

View publication stats