Materials Supply System Analysis Under Simulation Scenarios in a

Lean Manufacturing Environment

J. A. Jiménez-García*1, S. Téllez-Vázquez2, J. M. Medina-Flores3, H. H. Rodríguez-Santoyo4

and J. Cuevas-Ortuño5
1
Departamento de Ingeniería Industrial
Tecnológico Nacional de México
Instituto Tecnológico de Celaya
* [email protected]
2
Administración y Gestión de Pequeñas y Medianas Empresas
Universidad Politécnica de Guanajuato
3,4
Ingeniería en Energía
Universidad Politécnica de Guanajuato
Cortazar Guanajuato, México
5
Departamento de Ingeniería Industrial y de Sistemas
Tecnológico de Monterrey, campus Guadalajara
Guadalajara, Jalisco, México

ABSTRACT
Identifying the most efficient supply system for a company working under Lean Manufacturing practices was possible
with the support of this work. Promodel software was used to develop simulation model depicting a constant velocity
joints (CVJ) production system, where two different supply methods were assessed. According to results herein
obtained, better performance is achieved under random supply method in comparison with a clustering supply
method. The company’s goal is to keep 1% losses due to lack of material. In the actual process, this essential
parameter was reduced from 2.73% to 1.177%, if random supply method is properly implemented.

Keywords: Supply system, lean manufacturing, simulation.

"

1. Introduction

Supply chain management involves the Excess of raw material inventories and work in
coordination of several areas such as production, process are a major concern for an efficient supply
warehouse, location and transportation among its chain, and to mitigate their impact flow tools such as
members to achieve the best combination of pull system, SMED (single-minute exchange of
response capacity and efficiency to the market it died), balanced production, groups’ technology, etc,
supplies. There are five areas over which the are applied. However, in the extent the model
company can take defining decisions regarding its changes and process are decreased, the continuity
supply chain capacity, such as production, of cells is in risk due to lack of materials [3].
warehouse, location, transportation and
information [1]. This paper will focus on warehouse To counteract this effect, this work proposes to
and production, in the specific case of a constant- assess two supply methods in a Lean
velocity joints (CVJ) manufacturing company. Manufacturing-based company, performed with the
support of an outstanding too called Discrete-Event
The key strategy of supply chain is to be efficient Simulation (DES).
throughout the chain, and in order to achieve this
goal many worldwide companies are adopting Due to the nature of the company it is not possible
Lean Manufacturing production practices, which to modify actual facilities, since the company is
lead to a better competitive position [2]. After the classified as tier one within the automotive industry,
benefits obtained when implementing these i.e., it delivers the products directly to several car
practices, an evident problem arise such as manufacturers’ plants such as VW, Ford, Nissan,
losses due to a lack of material. Honda, GM, etc. To increase the rate for success in

Journal"of"Applied"Research"and"Technology 829
 Materials"Supply"System"Analysis"Under"Simulation"Scenarios"in"a"Lean"Manufacturing"Environment, J."A."JiménezどGarcía et"al."/"829ど838"

a Lean Manufacturing project, Standridge stands 2.4 Recommendations

that a convenient alternative is to apply simulation
[4]. It has been shown that the simulation can be After running the simulation model, it is possible to
applied to the analysis of supply chains [5]. This give recommendations based on the results.
paper proposes the use of the simulation using
Promodel software, giving flexibility to make 3. Production system description
experiments without using actual systems. Two
scenarios were executed in this model. In the first The manufacturing plant where this project was
one, the possibility of supplying the materials from developed produces constant-velocity joints (CVJ),
the warehouses to the production cells with a fundamental means of transmitting power from the
clustering approach is assessed, e.g., each forklift differential to the wheels. The production system
operator (router) can only deliver to the cells he was has 20 cells but only 17 are considered because of
assigned to. To cluster, the hierarchical clustering the demand of work; each one produces a family of
analysis (HCA) was used [6]. In the second products for different car manufacturers such as
scenario, any router can take materials from any Ford, Renault, Hyundai, VW, GM, etc. The
warehouse and deliver them to the production cell production quantities, the model and sequence in
that requires it. which they will be produced are recorded in the
heijunka box. This can be translated as production
Since the strategy of clustering has shown to give leveling or smoothing and within Lean
good results, a logistical problem of delivery and Manufacturing practices, it suggests to process
picking up of packages had been addressed, by small product lots on a frequent basis [8].
applying a strategy of sectorization, similar to the
one called cluster in our work [7]. Our proposal Figure 1 shows the cell 3; every cell has the same
compares the delivery and collection of materials configuration. Each cell counts with 4 operators who
using cluster against random deliveries. are responsible for the assembly of each one of the
components up to final product. They take all parts
The rest of the paper is structured as follows: that compose the product from a location called
section 2 defines the methodology; in section 3, the usage point. A resource called supplier boy, who
production system operation is described; section 4 puts each component in the usage point, avoids the
explains the materials supply system; in section 5, operators from moving in search of components.
the proposed supply systems are defined; in section
6 the steps to build the simulation model are listed; The finished goods are stored in racks called
in section 7 the results of both scenarios are shown; plastics, which purpose is to protect the finished
and in section 8, conclusions are given. goods. Once the amount of scheduled parts
indicated in the heijunka box is completed, a model
2. Materials and methods change is performed to begin the manufacturing of
the next scheduled product lot.
2.1 Description of the current production system

It is very important to understand the current

production system to define correctly the problem.

2.2 Description of the supply system

In the production system, there is a supply chain

system, that provides materials to the cells
production.

2.3 Simulation model

It is a platform that let us analyze the supply system

in different scenarios. Seven steps were followed
and will be described in section 6. Figure 1. Cell 3.

830" Vol."12,"October"2014"
 Materials"Supply"System"Analysis"Under"Simulation"Scenarios"in"a"Lean"Manufacturing"Environment, J."A."JiménezどGarcía"et"al."/"829ど838

4. Supply systems description model in the heijunka boxes on production cells.

The packs are composed by diverse materials
Wishing to process small product lots and in order (miscellaneous), which are parts like dampers,
to keep the parameter work in process (WIP) at a ties, locks, etc. For example, if a 240 piece lot is
minimum level a new supply system was scheduled for model x, the packer has to prepare
implemented. To fulfill the goal, a supply strategy the pack of miscellaneous composed by 480
was established, where materials are delivered just boots, 240 locks, 240 dampers, 240 ties, etc.
after the manufacturing of any model is completed
in the cell and the model change is triggered [3]. The supply boys attend two adjacent cells. Their
The materials handling system is composed by the role is to take materials that have been moved by
following resources: 5 packers, 3 routers (forklift routers and packers and to put them in the location
operators) and 9 supply boys assigned as seen on called usage point within the production cells, and
Table 1. In a simulation approach a resource from there operators take the parts and execute
means a person or equipment used to transport the corresponding assembly. This way the supply
raw material or finished good [9]. boys do not need to move from the cells to the
warehouse. This helps to avoid wasting great
RESOURCE ASSIGNED CELL amount of time due to unnecessary movements
Packer 1 C19, C20, and C21 and it allows the supply boy to attend up to two
Packer 2 C4, C2, C11, and C7 production cells [10].
Packer 3 C7, C15, C10, and C11
Packer 4 C12, C13, C14, C15, and 5. Supply based on clusters vs random supply
C17
Packer 5 C3, C6, C9, and C17 The way materials are supplied according to the
Router 1 C19, C20, C17, C9, C1 system is defined in Table 1. Such system was
and, C4 empirically designed, based on expertise from staff
Router 2 C10, C3, C5, C11, and members like supervisors, routers (forklift
C15 operators) and storekeepers. As it can be noted,
Router 3 C6, C12, C2, C7, C13, the scarcest resources are routers. They are
and C14 responsible for lack of materials on production
Supply boy 1 C19, and C20 cells, therefore the analysis will focus on them and
Supply boy 2 C2 and C4 improvements will be defined considering the
Supply boy 3 C1 and C9 opportunity to make modifications in the routers’
Supply boy 4 C6 and C12 operation. With regards to supply boys and
Supply boy 5 C11 and C5 packers, there will not be any modifications on
Supply boy 6 C13 and C7 their conditions.
Supply boy 7 C14 and C15
Supply boy 8 C3 and C10 5.1 Supply based on cluster
Supply boy 9 C17
Cluster is the classification of data in groups with
Table 1. Resources. similar features. The issue of clustering has been
approached in many situations by researchers in
The 3 routers are responsible for taking several sciences, and makes this tool very useful
materials from machined and plastics warehouse and easy to apply. Clustering is sometimes difficult
and move them towards the production cell that to develop due to its combinatorial complexity. Jain
has been assigned, just at the moment of model performed deep analysis of cluster techniques [6].
change for producing the next scheduled
product, with the purpose of having all materials Given the supply system characteristics, applying
available at the required time and thus to reduce the hierarchical clustering analysis (HCA) is here
WIP at minimum value. proposed, the HCA being usually used when the
exact number of clusters is required and few
The packers are responsible for assembling the elements are available. In this case, there will only
packs according to the scheduled amounts of each be three clusters, since there are only three routers

Journal"of"Applied"Research"and"Technology 831
 Materials"Supply"System"Analysis"Under"Simulation"Scenarios"in"a"Lean"Manufacturing"Environment, J."A."JiménezどGarcía et"al."/"829ど838"

to whom assign the different production cells that 6.1 Problem definition
they will attend.
It is one of the most important steps since if from the
5.2 Random supply beginning the project is properly defined, the
solution is obtained in great extent. The goal is to
Figure 2 shows the architecture proposed for reduce the losses (waiting) by lack of supply the
random supply. It can be detected that routers are production cells are subjected to when they are not
able to load materials from plastics and machined supplied in time, once the changeover is completed.
warehouse and deliver them to the production cell
as may be required. 6.2 Data acquisition and analysis

This means that any router can take materials from This stage consists in collecting data and
the plastics or machined warehouse and deliver defining the probability distribution they follow
them to the production cell that requires it, and are through goodness-of-fit tests. The most
not restricted to supply only to any given important data for building the model were the
production cell. The performance is assessed in following: router speed, distances from
the simulation model. warehouse to different production cells, each
cell’s cycle time, model changes duration and
6. Simulation model losses by lack of material. However, per request
of the company and its protection, the detail of
Due to the complexity of analyzing the system this step is not shown.
herein described, it is recommended to use
simulation. This tool has proven to be successfully 6.3 Model design
applied to projects related to Lean Manufacturing
practices, like Standridge concludes in his The simulation model was developed using
research [4]. The simulation model was built Promodel software and its huge capacity to depict
following a methodology of seven steps that will be discrete manufacturing systems. Figure 3 shows
outlined below [9]. the developed model layout.

Figure 2. Architecture of random supply system.

832" Vol."12,"October"2014"
 Materials"Supply"System"Analysis"Under"Simulation"Scenarios"in"a"Lean"Manufacturing"Environment, J."A."JiménezどGarcía"et"al."/"829ど838

Figure 3. Plant layout model developed with Promodel.

6.4 Verification and validation Validation was performed using two techniques
suggested by Harrell and Sargent: visual
Before taking any decision based on results validation and by comparison of historical
coming from the simulation model, it is records [10, 11]. The first technique implies
fundamental that the developed model exactly running the simulation process and to monitor
represents the actual system [9]. First, the the movements of resources such as packers,
verification must be made in order to assure the supply-boys and routers. It was also validated
model behaves as expected. Then validation is that cells started assemblies once they have
performed, which consists in defining if the materials like machined, miscellaneous, and
simulation model behaves just like the actual plastics. After it was detected that all movements
system. For the verification, the model was performed in the actual system are also included
developed in stages, i.e, elements were in the model, it was considered as validated by
progressively added and the model executed the visualization technique.
without being finished at all. This was made with
the purpose of detecting errors using the The second validation technique applied was
embedded debugging tool of the software. comparison of historical records from actual
Figure 4 shows a message box from the debug system against data coming from the simulation
tool. The message indicates that the model. Losses by lack of materials was the most
“miscelaneo” that would be sent to cell 6 has not important parameter, therefore it was analyzed in
yet declared the target location; therefore it is this study. Historical records were taken over the
not able to find its output. last eight months. As such, enough data could be
gathered since 20 data can be collected every
" " " X" month. To validate the results of the simulation
Translation"Status:"Normal"Run"
" " " " "
" " " "
model, we used a 95% confidence intervals on a
Process"and"Routing" difference in mean to define if average losses by
For"MISCELANEO_L6"at"PRESURTIMIENTO_L6"
Routing"Block"#1" lack of materials in the real system is equal with
Route"#1"Output"entity:"MISCELANEO_L6" the results of the simulation model, using
ERROR:"A"process"record"has"not"been"defined"for"the"output"
entity"at"the"destination"location!" equation (1) [12]:
"
"
な な
" 博な 伐 捲博に 伐 建糠 券な髪券に伐に 鯨喧 謬
捲 髪 博な 伐
判 航な 伐 航に 判 捲
" " " Abort" " Continue" " Detailed"Status に 券な 券に
" " " " " " " " " " "
" " " " な な (1)
博に 髪 建糠 券な髪券に伐に 鯨喧 謬
捲 髪
に 券な 券に
Figure 4. Software debugger tool to verify model.

Journal"of"Applied"Research"and"Technology 833
 Materials"Supply"System"Analysis"Under"Simulation"Scenarios"in"a"Lean"Manufacturing"Environment, J."A."JiménezどGarcía et"al."/"829ど838"

We assume that the sample means and variances achieved when all manufacturing cells had started
are from two independent normal populations. We at least their first scheduled production lot. Taking
did a goodness-of-fit test to ensure that the data this into consideration, a variable which counts
follow a normal distribution. Figure 5 shows an from zero was implemented to the model, “zero”
example with cell 12. Every set of data was tested meaning that no cell has started to operate, up to
in order to validate the normality. 17, meaning that all production cells had started to
process at least their first scheduled lot. 15
"
Cell 12
repetitions were executed to define the time at
Normal which a steady state is achieved and it was found
99
Mean 3.186 that in each repetition, the variable reaches the
StDev 1.408
95 N 45 value of 17 after 4 hours. This allowed defining a
AD 0.362
90
P-Value 0.428 warm-up period of 4 hours.
80
70
Percent

60
50 CONFIDENCE INTERVALS
40
ON A DIFFERENCE IN
30
SIMILAR
20 CELL MEAN
MEAN?
10 ACTUAL SYSTEM VS
5
SIMULATION
1 1 -1.756, 0.738 Yes
0 1 2 3 4 5 6 7
L12
2 -1.221, 0.868 Yes
3 -1.882, 0.377 Yes
4 -0.473, 0.954 Yes
Figure 5. Goodness-of-fit normality in data of cell 12. 5 -0.361, 1.083 Yes
6 -1.331, 0.771 Yes
Results from the simulation model and historical 7 -0.451, 0.752 Yes
data from current system were considered as well 9 -0.554, 0.524 Yes
as 95% confidence interval on a difference in 10 -2.907, 0.070 Yes
means for the losses due to lack of material 11 -0.748, 1.155 Yes
coming from the simulation model and current 12 -1.612, 1.827 Yes
system. The confidence intervals are listed on 13 -0.040, 0.860 Yes
Table 2, where only 17 cells currently working 14 -1.185, 0.030 Yes
according with the demand appear. 15 -1.826, 0.736 Yes
17 -0.399, 0.481 Yes
If the confidence interval includes zero, it implies 19 -0.036, 1.106 Yes
that the means are equals [12]. In the results 20 -0.021, 0.840 Yes
listed in Table 2, it can be detected that every
confidence interval includes zero; therefore, the Table 2. Model validation.
means of the simulation model and actual system
are equal and that the simulation model is 6.6 Experimental
statistically considered valid and can be used as
a tool to perform the experiment. Two experiments were performed. In the first
one the model was programmed considering the
6.5 Warm up determination clusters defined in Figure 7. In the second
experiment the model was programmed under
In a simulation model, it is important to define the random supply design, in which the restriction
time at which steady state conditions are met [13]. for routers to deliver only to its assigned cluster
The statistical information collected during the was removed.
warm-up stage is not considered for analysis, and
therefore Promodel software discards the 6.7 Results analysis
information created during such period. The
simulation starts when the system is empty, which The results are described in detail in the following
is not a steady state condition. This can only be section.

834" Vol."12,"October"2014"
 Materials"Supply"System"Analysis"Under"Simulation"Scenarios"in"a"Lean"Manufacturing"Environment, J."A."JiménezどGarcía"et"al."/"829ど838

7. Results The information in Table 3 was collected with

Minitab statistical software, to define clusters using
To define clusters based on HCA, the distances the following steps: Stat – Multivariate – Cluster
from each production cell regarding the warehouse Observations – Variables or distance matrix – Link
were defined, marked with position (0, 0) as can be method (Average) - Distance measure (Squared) –
seen in Figure 6, where the distances are in Number of cluster (3) – Show dendogram –
meters. The distances were considered quadratic, customize – Title (Clusters) - Case labels (Cells) –
because of the limits of the configuration. Label Y axis with (Distance) – Show dendogram in
(One Graph) – ok.

After completing the steps previously mentioned,

the final clustering is clearly seen in Figure 7. This
clustering suggests that Router 1 supplies C1, C2,
C3, C4, C5, C9, and C10 cells (in red). The router
2 supplies C7, C11, C13, C14, and C15 cells (in
blue); and finally, the router 3 supplies C6, C12,
C17, C19 and C20 cells (in green).

CLUSTERS

5.41
Figure 6. Plant layout.

Considering distances shown in Figure 6, it is 3.61

Distance

possible to define the location of cells in a

Cartesian plane in the first quadrant and then
record their X and Y values, which are listed in 1.80

Table 3.
0.00
C1 C2 C9 C4 C3 C10 C5 C7 C11 C13 C14 C15 C6 C17 C12 C19 C20
LOCATION Observations

CELL X Y
C1 51 35 Figure 7. Clusters obtained with Minitab.
C2 58 33
C3 51 48 This proposal will be analyzed on the simulation
C4 42 35 model to calculate its performance before
C5 34 48 implementing it to the actual system. Each
C6 35 20 repetition was executed during 24 hours, since
C7 26 35 statistics obtained from actual system were
C9 58 33 obtained considering this period of time. The
C10 60 48 following table lists the results when routers supply
C11 24 38 materials by HCA clustering. Using the tool Stat:fit,
C12 26 20
it was defined that for a 95% confidence level and
1% of maximum allowed error, 50 repetitions were
C13 15 35
required. The average results are listed in Table 4.
C14 5 35
C15 5 48 For the random supply scenario and also
C17 41 20 considering a 95% confidence level and 1% of
C19 60 17 maximum allowed error, it was defined that for
C20 50 17 proper average calculation, executing 45 repetitions
was required. The results are listed in table 5.
Table 3. Cells location.

Journal"of"Applied"Research"and"Technology 835
 Materials"Supply"System"Analysis"Under"Simulation"Scenarios"in"a"Lean"Manufacturing"Environment, J."A."JiménezどGarcía et"al."/"829ど838"

PERCENTAGE Table 6 lists a summary with results, where the

ASSIGNED OF LOSSES BY average system performance for each scenario
CLUSTER ROUTER
CELL LACK OF can be seen.
MATERIAL
1 C1 4.521
1 C2 3.59 PERCENTAGE OF
SUPPLY
1 C3 6.244 WAITING BY LACK
1 SYSTEM
1 C4 4.632 OF MATERIAL
1 C5 3.473
ACTUAL 2.71
1 C9 5.327
1 C10 5.505 CLUSTERS 3.51
2 C7 2.616
2 C11 3.613 RANDOM 1.177
2 2 C13 1.307
2 C14 3.418
Table 6. Losses by lack of material for
2 C15 5.002
current, clusters and random supply systems.
3 C6 2.484
3 C17 4.037
3 3 C12 2.639 8. Conclusions
3 C19 0.739
3 C20 0.573 Three different scenarios were studied, current
AVERAGE 3.513 supply system, supply system using clusters defined
by HCA method and random supply method.
Table 4. Losses by lack of material for current According to the results listed in Table 6 from the
clusters and random supply systems. previous section, it is recommended to use the
random supply method, taking into account that the
PERCENTAGE OF goal is to reduce losses by lack of material.
CELL WAITING BY LACK
OF MATERIAL The random supply system is very complex, since
C1 1.85 any router (the first that is available) can take
materials from any warehouse (plastics or
C2 1.24
machined) and deliver them to the production cell
C3 2.02 that is making its request. To decrease complexity,
C4 1.14 it is recommended to combine the system with an
C5 0.33 Andon system, which has visual and audio
C6 2.03 elements helping to inform about downtimes under
certain circumstances. In this case, a blue color is
C7 0.16
suggested to point out that materials are required
C9 1.54 since most authors suggest the blue color for the
C10 0.88 lack of material.
C11 0.73
C12 2.18 The Andon system should be synchronized to
bond the cell needs with the first available router
C13 0.23
since it was able to implement it in the simulation
C14 0.88 process. However, to apply it in the actual system
C15 1.26 would be hard to accomplish, but not impossible.
C17 3.14
C19 0.30 To make the final decision, several criteria must be
C20 0.10
considered. For example, if clustering supply
method is used, the materials supply management
AVERAGE 1.177
is made easier, in great extent due to the fact that
Table 5. Losses by lack of material every router knows in advance the cells that will be
under random scenario. supplied, but it faces losses by lack of materials.

836" Vol."12,"October"2014"
 Materials"Supply"System"Analysis"Under"Simulation"Scenarios"in"a"Lean"Manufacturing"Environment, J."A."JiménezどGarcía"et"al."/"829ど838

On the other hand, if random supply is selected, References

the milestone will be that different routers deliver at
the exact moment to the production cells, i.e, when [1]-M. H. Hugos. “Essentials of Supply Chain
they need it. management”, third edition. United States of America.
John Willey, 2011. pp. 21-26.
Using this proposal, it makes it easier to apply [2] D. Rajenthirakumar, and R. Ghowtam Shankar,
flow tools related to Lean Manufacturing, like “Analyzing of the benefits: a consumer durables
small production lots, to make frequent model manufacturing company case study,” International Journal
changes, balanced production, WIP reduction, of engineering. Annals of Faculty Engineering Hunedoara.
etc., since materials will be delivered at the London, Tome IX, Fascicule 3. pp. 335–339, 2011.
moment they are required.
[3] J. A. Jiménez et al. “Diagnostico en el surtimiento de
An important restriction for applying this solution is materiales en líneas de ensamble de flechas de
velocidad constante Utilizando Promodel,” En: Actas del
that company’s production must be performed 1er Congreso Internacional de investigación e
under Lean Manufacturing practices. Besides, it is innovación 2012. Multidisciplinario, (26 y 27 de abril de
required that staff is capable to make simulations, 2012). ISBN: 978-607-95635.
in order to first analyze the supply of materials in
the simulation model. [4] C. R. Standridge and J. H. Marvel. “Why lean needs
simulation”. Proceedings of the 2006 Winter Simulation
In the three supply systems; actual, clustering and Conference. p.1907-1913. ISBN: 1-4244-0501-7.
random, the routers can only take one load and
deliver it to its location (from the warehouse to the [5] M. Cedillo and C. Sanchez. “Dynamic Self-Assessment
of Supply Chains Performance: an Emerging Market
cells for the case of raw materials or from the Approach”, Journal of Applied Research and Technology.
production cell to safe launch area in the case of Vol. 11, June 2013, pp. 338-347.
finished Goods). Therefore it is recommended to
analyze scenarios where the possibility for routers [6] A. K. Jain et al. “Data clustering: a review”, ACM
to deliver more than one load and make materials Computing Surveys. Septiembre 1999. Vol. 31, No.3.
deliveries in one single trip, performing loads and p.264-323. DOI: 10.1145/331499.331504.
unloads, delivering raw materials and taking
finished goods is considered. For further research, [7] R. G. González et al. “A Hybrid Metaheuristic
modifying the company’s supply system is Approach to Optimize the Districting Design of a Parcel
Company”, Journal of Applied Research and
proposed, to apply the mentioned restrictions in Technology. Vol. 9 April 2011, pp. 19- 35.
the vehicle router problem (VRP) where more than
one load can be taken and in a single trip load and [8] J. W. Ohlmann, et al. “Route Design for Lean Production
unload materials, without the need of returning to Systems”. Transportation Science. Agosto 2008. Vol. 42.
the warehouse. No.3. p.352-370. DOI:10.1287/trsc.1070.0222.

Acknowledgments [9] C Harrell, B. K. Gosh y R Bowden. “Simulatión Using

Promodel”, 3a ed. United States of America: Mc Graw
This project was performed thanks to support given by Hill Higher Education, 2000. pp. 194-217.
Conacyt through scholarship No. 228973 for funding my
specialization studies and we wish to thank CIATEC, [10] J. A. Jiménez et al. “Propuesta de mejora basada
where training about industrial engineering topics were en manufactura esbelta para el surtimiento de
given by expert personal. materiales en un célula de ensamble utilizando la
simulación” Academia Journals. Noviembre 2012. Vol. 4
No. 3 Tomo 11, pp. 1447-1452. ISSN 1946-5351.

[11] R. G. Sargent. “Verification and validation of

simulation models”. Proceedings of the 30th conference
on Winter simulation. Department of Electrical
Engineering and Computer Science. 1998. p.121-130.
ISBN: 0-7803-5134-7.

Journal"of"Applied"Research"and"Technology 837
 Materials"Supply"System"Analysis"Under"Simulation"Scenarios"in"a"Lean"Manufacturing"Environment, J."A."JiménezどGarcía et"al."/"829ど838"

[12] D. C. Montgomery and G. C. Runger. “Applied

statistics and probability for engineers”, Third edition.
Arizona: John Willey, 2003. pp. 250-290.

[13] E. J. Chen and W. D. Kelton. “Determining

simulation run length with the runs test”. Simulation
Modelling Practice and Theory. Volume 11. July 2003. "
pp. 237-250. "
"
"
"
"
"
"
"
"

838" Vol."12,"October"2014"