08MScLIXPU1 2

INVESTIGATION OF DYNAMIC
PROCESS PLANNING USING

COMPUTER SIMULATION ARENA

By

Pramote Umnutkittikul

MSc Operations Management

Investigation of Dynamic Process Planning
using Computer Simulation Arena

By

Pramote Umnutkittikul

2008

A Dissertation presented in part consideration for
the degree of MSc Operations Management

Abstract
Due to the dynamic fluctuation of global market and customer demands, manufacturing firms are
facing difficulties in rapidly responding to market changes. Many firms are under significant
pressure to produce low cost, high quality products that can compete in the world marketplace. It
might be useful if the process plan could be dynamically modified to consider the current state in
the manufacturing shop floor. Thus, the concept of integration of process planning and
scheduling function has been established so as to help the firms to handle an unstable market.
This dissertation is prepared to investigate the dynamic process planning according to the
integration of process planning and scheduling by the use of computer simulation. By using
Arena software, the models have been developed to replicate the actual manufacturing system.
They are presented either in static or dynamic manner. To start the experiment, a set of machines
and jobs along with processing times are required from the actual data of Thailand based
company. To monitor the manufacturing performance in dynamic process plan, the relationship
of alternative machines has been applied. The research mainly observes the relationship between
numbers of product being produced, production time, waiting time, number of item waiting and
also instantaneous use of machines in the job shop condition.

Table of Contents

CHAPTER 1 - INTRODUCTION 1
1.1 Aims and Objectives of the Dissertation 1
1.2 Trends of Market 2
1.3 Need for Integration 4
1.4 Overview of the Dissertation 6

CHAPTER 2 LITERATURE REVIEW 7
2.1 Process Planning 7
2.1.1 Computer-aided Process Planning System 8
2.2 Scheduling Function 10
2.2.1 J ob Shop Scheduling 11
2.2.2 Dispatching Rule for Dynamic Scheduling 12
2.3 Integration of Process Planning and Scheduling Function 14
2.3.1 Approaches of the Integration 15
2.3.2 Classification of the Integration Approaches 18
2.4 Computer Simulation 21
2.5 Summary of the Research 23

CHAPTER 3 RESEARCH METHODOLOGY 25
3.1 Data for Simulation Experiment 25
3.2 Simulation Model Building 27
3.2.1 Static Model using Sequence Method 27
3.2.2 Static Model using Variable Method 32
3.2.3 Dynamic Model 36
CHAPTER 4 EXPERIMENTS & RESULTS ANALYSIS 39
4.1 Experiment 1:
Comparison of Static Model using Sequence Method and Static
Model using Variable Method 39
4.2 Experiment 2:
Comparison of Static Model and Dynamic Model 42
4.3 Experiment 3:
Effects to Manufacturing System after Changing the Replication Length 45
4.4 Experiment 4:
Effects to Manufacturing Performance as the Number of Alternative
Routes Increasing 47

CHAPTER 5 DISCUSSION 51
5.1 Discussion of Experiment 1 51

CHAPTER 6 CONCLUSION 57
6.1 Research Conclusion 57
6.2 Recommended Further Work 59

LIST OF REFERENCES 60

APPENDICES 66
APPENDIX A
Comparison of Integration Models 67
APPENDIX B
Static Model Using Sequence Method 68
APPENDIX C
Static Model Using Variable Method 70
APPENDIX D
Dynamic Model 72
APPENDIX E
Table of Alternative Routes for Experiment 2 & 3 75
APPENDIX F
Table of Alternative Routes for Experiment 4 76
Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Chapter 1
Introduction

1.1 Aims and Objectives of the Dissertation
To investigate the dynamic process planning regarding to the integration of process
planning and scheduling function using computer simulation.
To develop simulation model by using Rockwell Software Arena. The models are
presented either in static or dynamic manner.
To carry out the simulation experiments so as to observe the manufacturing performance
of the system in different circumstances
To improve the manufacturing performance according to the alternative routes approach.
In this case, alternative machines concept has been applied.
In General, the manufacturing process is planned in a static manner, whether it is prepared by
human or with computer assistance. However, due to the dynamic fluctuation of customer
demands in the market, manufacturing enterprises are facing difficulties in rapidly responding to
market changes (Zhang, 2004). Thus, it might be useful if the process plan could be dynamically
modified to consider the current state of the manufacturing system so as to study the
manufacturing performance in unpredictable situation. For the purpose of increasing the
responsiveness of manufacturing systems to handle unstable market changes, the integration of
process planning and production scheduling concept has been introduced.
Process planning is normally concerned with the selection of various operations/processes to be
performed in sequence to convert a part from its initial form to final form according to
specification (Sundaram and Fu, 1988). On the other hand, scheduling is the allocation of the
resources in the shop over time to manufacture the various parts (Zhang et al, 2003). However,
1

these two systems often treat an incoming job without paying much attention to the effect which
changing shop-floor conditions. Consequently, Zhang et al (2003) suggested that the process
plan generated from such systems are not entirely followed in shop-floor implementation due to
resource unavailability, machine breakdown, bottleneck and so on. Therefore, it can be said that
the process plans should be flexible in order to manage the dynamic changes on the shop floor
(Wongthanasunthorn, 2007). In addition, the resulting work schedule should be able to reflect
this flexibility and respond with the changes.
In this dissertation, the overview of process planning and scheduling function together with
background of the integration of those two functions are firstly presented. After that, the
simulation model will be built so as to replicate the shop floor environment and inspect the
manufacturing performance during the experiments. This study also intends to compare the
results in terms of performance obtained when considering either a single fixed route approach or
an approach of alternative routes. According to the simulation experiment, number of product
being produced, production time, waiting time, number of item waiting and machine utilization
are mainly focused either in static or dynamic situation.

1.2 Trends of the Market
In recent years, manufacturing enterprises have been under pressure to competently cope with a
market that is rapidly changing due to global competition, shorter product life cycles, dynamic
changes of demand patterns and product varieties (Lim and Zhang, 2004). The demand of
manufacturing products is becoming more and more oriented to high variety and quality products
(Carvalho, 1996). Consequently, customer satisfaction has become tougher to achieve and
manufacturing firms have to be agile and responsive to market changes in order to stand firm in
the competitive market (Hormozi, 1994). Improving manufacturing productivity basically
presents a critical challenge to most production managers.
The changes in market demands, product life and technological achievements, specifying the
new environment as illustrated by Larsen and Alting (1992) in figure 1.1:
2

Figure 1.1: A figure of changes in environment nowadays <Larsen and Alting, 1992>

Lim and Zhang (2004) stated that manufacturing systems are expected to be able to achieve
shorter lead-times and better quality in products and delivery. In addition, manufacturing systems
are also required to be adaptive and responsive to cope with not only external turbulence, such as
dynamic market changes and technological changes, but also internal uncertainties (e.g. resource
shortages, machine breakdowns and etc.). Hence, the manufacturing companies must acquire
abilities to respond quickly to rising demand of products with better quality at lower price in
order to successfully compete in the business world.
In order to achieve this rapid reaction to the changes in demand manufacturing, Carvalho (1996)
proposed that systems need to perform many of the production tasks in an integrated manner.
The concept of concurrent engineering assumes an information sharing environment to support
all phases of product development, planning, production scheduling and manufacture (Carvalho,
1996). These efforts must be directed towards eliminating the division between process and
production planning activities to achieve an integrated manufacturing environment in which
3

business processes give due recognition to the impact of their decisions on related manufacturing
activities (Carvalho, 1996). Although the literature of Shao et al (2008) stated that there is a
strong relationship between process planning and scheduling, the integration of them is still a
challenge in both research and applications. Overall, it is believed that improvement of the
interaction and integration between process planning and scheduling could lead to a dynamic
manufacturing environment. The integration is capable of reacting to a real-time shop floor
conditions, and can significantly reducing the cost and lead-time as well as improving
manufacturing responsiveness.

1.3 Need for Integration
According to the research of Carvalho (1996), manufacturing companies have been functionally
divided into separate specialist departments performing computer aided design (CAD),
computer-aided process planning (CAPP) and production planning and control (PPC). In
traditional approach, process planning and scheduling were carried out sequentially, where
scheduling was done separately after the process plan had been generated (Shao et al, 2008).
Process planning typically identifies machining processes, resources and parameters necessary to
convert the raw materials into finished products (J oo et al, 2001). Scheduling plans receive
process plans as their input and their task is to schedule the operations on the machines while
satisfying the precedence relations given in the process plans (Nahmias, 2005). Nevertheless,
these approaches have become an obstacle to enhance the productivity and responsiveness of
manufacturing systems, and they may cause the following problems (Shao et al, 2008):
1) In a manufacturing organization, process planning function works in static. It considers
the resources on the shop floor in an ideal way. Process planners assume unlimited
resources on the shop floor and plan for the most recommended alternative process
(Usher and Fernandes, 1996). This may lead to the process planners favoring to select the
desirable machines repeatedly. Therefore, the generated process plans are somewhat
unrealistic and cannot be readily executed on the shop floor (Lee and Kim, 2001).
4

Accordingly, the resulting optimum process plans often become infeasible when they are
carried out in practice at the later stage.
2) Even if, in the planning phase, process planners consider the current resources on the
shop floor, the constraints considered in the process planning phase may have already
changed greatly because of the time delay between planning phase and execution phase.
This may lead to the optimized process plan infeasibility. Investigations have shown that
20-30% of the total production plans in a given period have to be modified to adapt to the
dynamic changing of a production environment (Kumar and Rajotia, 2003).
3) Scheduling plans are often determined after process plans. In the scheduling phase,
scheduling planners have to consider the determined process plans. Fixed process plans
may drive scheduling plans to end up with severely unbalanced resource load and create
superfluous bottlenecks.
4) In most cases, both for process planning and scheduling, a single criterion optimization
technique is used for determining the best solution. However, the real production
environment is best represented by considering more than one criterion simultaneously
(Kumar and Rajotia, 2003). Furthermore, the process planning and scheduling may have
conflicting objectives. Process planning emphasizes the technological requirements of a
task, while scheduling involves the timing aspects of tit. If there is no appropriate
coordination, it may create conflicting problems.
To overcome these problems, there is thus a major need for an integrated process planning and
scheduling system (Shao et al, 2008). Moreover, the way to successfully achieve good results
from the integration between process planning and scheduling was suggested by Carvalho (1996)
as it is necessary to identify the possible alternative plans and select the ones that will better
perform once the products are being manufactured. The integration of the two functions may
introduce significant improvements to the efficiency of the manufacturing facilities through
elimination or reduction in scheduling conflicts, reduction of flow-time and work-in-process,
improvement of production resources utilization and adaptation to irregular shop floor
disturbances (Lee and Kim, 2001).
5

1.4 Overview of the Dissertation
This dissertation is divided into six chapters including introduction chapter which identifies the
aims and objectives of the research in company with the trends of the market and important of
the integration between process planning and scheduling function.
Chapter 2 briefly reviews the literature and background of the related work in the areas of
process planning and scheduling function along with the integration of these two functions.
Moreover, the overview of the computer simulation is described in this chapter.
Chapter 3 illustrates the resource data obtained from the actual company which used in the
experiments and then introduces the procedures to build the simulation models. The models are
simply created using Rockwell software Arena. The models contain two static models and one
dynamic model.
Chapter 4 provides the examination of different simulation experiments. The experimental
results are also analyzed in this chapter.
Chapter 5 presents the comment and discussion of the results obtained from the previous chapter.
It is divided into four topics according to the experiments.
Finally chapter 6 draws the conclusion of the dissertation as well as proposes some further work
to be made in this area.

6

Chapter 2
Literature Review

This chapter illustrates the summary of previous works and literatures associated to the research
topic. It begins with the background knowledge of process planning and scheduling
independently. Then, the integration between process planning and scheduling in manufacturing
systems will be described along with the texts on the issue of computer simulation.

2.1 Process Planning
Shao et al (2008) indicated that a process plan specifies what raw materials or components are
needed to produce a product, and what processes and operations are necessary to transform those
raw materials into the final product. Process planning typically finds the best way to manufacture
the part by finding the optimal processing sequence of the features and the ideal machines to
process the part (Chen and Khoshnevis, 1993). Its outputs include manufacturing processes,
process parameters and identification of the required machines, tools and fixtures needed to
perform those processes (Dong et al, 1992). The ability of manufacturing to respond; to changes
in product design, introduction of new products and changes in resource capabilities, depends
largely on process planning. Process planning is therefore vitally important in providing a
competitive edge to manufacturing (Bhaskaran, 1990).
Carvalho (1996) noted that process planning is traditionally performed by experienced planners
who based on the characteristics of the product; decide on all the necessary operations and
equipment (resources) needed to produce the required product. The decision of process planner
will become the production routing that will be used by production engineer for capacity
planning, facility planning, costing and etc. (Halevi, 1980). Nonetheless, the main drawbacks of
the manual process planning are that it is very time-consuming and there is no consistency in the
decision making by humans. On many cases, one planner would create one specific plan for the
7

product, whereas the others may create dissimilar plans in different circumstances so as to
manufacture exactly the same product. Furthermore, process planners often overlook some
essential factors such as the quantities, timing, availability of the resources and capabilities
(Wongthanasunthorn, 2007).

2.1.1 Computer-aided Process Planning System
Over the last two decades, computer-aided process planning (CAPP) has received a widely
research attention from many researchers. Workers currently focus more on their ability to
program and monitor numerical-controlled (NC) machine tools rather than make decisions about
sequencing or selection machine operations in order to produce a component (Carvalho, 1996).
This tool enables the conversion of the geometric models into machine code which represents a
machine sequence. Groover and Zimmers (1984) also pointed that the CAPP system focuses on
1) the creation of part specification data from a coding scheme, 2) accessing a standard routing
file for machine assignments, and 3) the translation of part specifications to NC instruction code
for machine operations. CAPP systems appear to be capable of creating more accurate and
consistent plans in shorter time with reduction in skilled process planner requirements (Carvalho,
1996). As a result, it can simplify and improve process planning to achieve more effective use of
manufacturing resources. Although it has not been easy to put all the process planning
experience collected by humans to computer, the use of computers to generate process planning
is growing extensively in manufacturing industry. Chang (1990) illustrated that computer-aided
process planning can help in the planning decision and has a lot more advantages:
It reduces the demand on the skilled planner.
It reduces the process planning time.
It reduces both process planning and manufacturing cost.
It creates consistent plans.
It produces accurate plans.
8

It increases productivity.
From the flexible manufacturing systems (FMS) view point, most current CAPP systems are not
able to take advantage of the flexibility offered by FMS, because they were designed in such a
way that it usually generates only one process plan for each given part with only one routing
and one set of machines (Dong et al, 1992). Furthermore, the current CAPP system usually
selects and recommends a preferred manufacturing process based on technological and/or
economic considerations. Hence, some machines are heavily loaded which made them become a
bottleneck, while others are under-utilized.
Two main approaches can be defined to the CAPP: variant and generative by Carvalho (1996).
The very first CAPP systems developed were based on the variant approach while nowadays the
effort is more noticed on the generative approach. The explanation of these two approaches is
expressed below:
Variant approach: This approach operates the concept of group technology. It assumes
that parts from the same family have similar process plans. A standard process plan is
generated for each part family and stored in a database (Carvalho, 1996). This approach
is useful when the part families are grouped by their similarities in terms of processes
(Chen et al, 1990). The majority of the most successful application of CAPP systems in
industry is variant approach (Gu and Zhang, 1994). Some examples of systems based on
the variant method are MIPLAN by Alting and Zhang (1989), MULTIPLAN by J oshi et
al (1988) and etc.
Generative approach: It synthesizes a new plan for each part (Carvalho, 1996). A system
is built where the rules and decision algorithms establish the link between the
characteristics of component and machine capability requirements (Carvalho, 1996).
Moreover, such systems also optimize the selection of available tools, fixtures, machines,
sequence of operations and so on. Some notable examples are KAPLAN (Giusti and
Santochi, 1989), ACES (Wong and Siu, 1992), GENPLAN (Gindy et al, 1993) and etc.
Although several CAPP systems have been reported in recent studies, few of them have been
practically applied due to their lack of the ability to cope with the dynamic nature of
9

manufacturing systems (Ming et al, 1997). The implementation of such systems in
manufacturing companies did not happen as expected (Ali and Motavalli, 1993). A major reason
for the shortcomings of the current process planning practice is the unrealistic assumptions.
Those assumptions are identified by Larsen and Alting (1992) as:
Process plans are assumed to be static and fixed because they are produced in absence of
dynamic information. Since no alternatives are built into the static process plan, the
process plan information is very constrained.
Unlimited shop floor resources (materials, machines, tools, fixtures, operators, etc.)
availability is assumed during process planning. But usually they are engaged with other
tasks, such as processing, breakdown, maintenance, absence, etc.
The desirable resources are repeatedly selected during planning, although the only reason
is the process planners personal preference.

2.2 Scheduling Function
Scheduling is a decision-making process that plays an important role in most manufacturing and
service industries (Pinedo and Chao, 1999). With increase emphasis on time to market and time
to volume as well as improved customer satisfaction, efficient scheduling will gain increasing
emphasis in the operations function (Nahmias, 2005). Scheduling is treated as a time dependent
function aimed at the utilization of resources to satisfy the process plans for a number of
products (Chryssolouris and Chang, 1985; Kulnle et al, 1994). Generally, scheduling and
sequencing is concerned with loading jobs on machines so that the available resources can be
utilized effectively (Sundaram and Fu, 1988).
Carvalho (1996) classified the scheduling method as static or dynamic according to the job
availability. It is static if it is applicable when all jobs are available in the beginning and dynamic
if it is applicable when jobs arrive continuously. In another way, a scheduling also can be
classified as either real-time or off-line (Carvalho, 1996). In a real-time, schedules can be
changed when an event occurs while in an off-line method; the schedules are generated in
10

advance and then strictly followed during the manufacture of the products in an off-line
scheduling (Carvalho, 1996). As the scheduling algorithm assumes a given set of jobs with fixed
process plans and generates a good schedule according to some criteria, Zhang et al (2003) thus
argued that this could often lead to schedules with severely unbalanced resource loading and
bottleneck machines, which in turn lead to lower overall resource utilization and poor delivery
performance.

2.2.1 Job Shop Scheduling
Carvalho (1996) defined a job shop scheduling as a manufacturing facility which makes use of
universal resources and can be used for variety of manufacturing operations. Incoming parts
normally flow through the job shop in a multiple routes according to their manufacturing
requirements. There are many different types of scheduling problems faced by the firm including
personnel scheduling, job shop scheduling, facilities scheduling, vehicle scheduling and etc.
(Nahmias, 2005). J ob shop scheduling, which is focused in this project, is known more
commonly in practice as shop floor control. In fact, Nahmias (2005) classified the job shop
scheduling problem based on five important characteristics including job arrival pattern, number
and variety of machines in the shop, number of workers in the shop, particular flow patterns and
evaluation of alternative rules.
Studies about J ob shop scheduling problem have been carried out for the last sixty years and
there is still a big gap between schedules generated through various planning systems and
schedules actually use in the shop floor environment (Dagli and Sittisathanchai, 1993).
Bradimart (1993) identified the job shop scheduling problem as one of the hardest discrete
optimization problems. It is widely known that the job shop scheduling problem can be
formulated as follows: consider n jobs and m different machines and each job consists of a
number of operations o which have to be processed in a certain order and each one of those
operations must be carried out on a certain machine (Carvalho, 1996). The number of possible
schedules is a total of (n!)
m
and each of them must then be examined to select the one that gives
the best performance (Rembold et al, 1993). This kind of scheduling problem is considered to be
11

a NP-hard problem, which an optimal solution is unlikely to be found. When the time required
finding the optimal solution increases exponentially, the size of NP-hard problem is increasing as
well (Rodammer and White, 1988).

2.2.2 Dispatching Rules for Dynamic Scheduling
Many methods have been used to produce schedules but heuristic methods have shown
significant promise in solving the general job shop scheduling problem (Carvalho, 1996). The
most common heuristics belong to the class of procedures called dispatching rules. Pinedo and
Chao (1999) defined a dispatching rule as a rule that prioritizes all the jobs that are waiting for
processing on a machine. The prioritization scheme may take into account the jobs attributes
and the machines attributes, as well as the current time. In the field of sequencing and
scheduling research, most of the theoretical work has been applied to static problems, where the
data about all jobs to be processed is know from the start (Carvalho, 1996) However, a more
realistic challenge is the dynamic scheduling problem, where job characteristics are not known
until they arrive at the shop floor (Carvalho, 1996). Research in dispatching rules has been
examined for several decades, and many rules have been developed and studied. The following
dispatching rules are among the most common rules used in practice and found in the texts:
First-come, first-served (FCFS): J obs are processed in the sequence in which they
entered the shop (Nahmias, 2005). This is known as the most basic dispatching rule
which assure that the jobs will not waiting for a long time in queue.
Shortest processing time (SPT): J obs are sequenced in increasing order of their
processing times. The job with the shortest processing time is first, the job with the next
shortest processing time is second and so on. For a single machine system, this rule
optimizes average flow time and average lateness (Carvalho, 1996). Conway and
Maxwell (1962) found that within multiple-machine environment this rule retains the
advantages throughput maximization which it has exhibited in the single server
environment.
12

Earliest due date (EDD): J obs are sequenced in increasing order of their due dates. The
job with the earliest due date is first, the job with the next earliest due date is second and
so on. Pinedo and Chao (1999) noted that this rule tends to minimizes maximum lateness
and maximum tardiness among the jobs waiting for processing. EDD normally performs
better than SPT at high allowances at minimizing average tardiness (Carvalho, 1996).
Critical ratio scheduling (CR): After each job has been processed, the critical ratio ([Due
date Current time] / Processing time) has been computed and the next job will be
scheduled in order to minimize the value of the critical ratio (Nahmias, 2005). The idea
behind critical ratio scheduling is to provide a balance between SPT, which only
considers processing time and EDD, which only considers due dates. One disadvantage
of the method is that the critical ratios need to be recalculated each time a job is
scheduled (Nahmias, 2005).
In consideration to Pinedo and Chao (1999) literature, the basic dispatching rules described
above are of limited use. When a complex objective has to be minimized, none of the basic
dispatching rules may perform effectively. Thus, Pinedo and Chao (1999) suggested that a
framework for combining basic dispatching rules can make the resulting rules perform
significantly better.
Apart from the dispatching rules, there are a numerous methods used to solve the problem of job
shop scheduling including enumeration methods, artificial intelligence (AI), integer
programming, rescheduling method, neural network structure and so on. Enumeration method,
for instance, basically lists all possible schedules and then eliminates the non-optimal
possibilities from the list, leaving those which are optimal (Carvalho, 1996). For artificial
intelligence approaches, they typically depict the scheduling problem as the determination and
satisfaction of the large number and variety of hard and soft constraints that are found in the
scheduling domain (Carvalho, 1996). AI is used to extend knowledge representation techniques
to capture these constraints, to integrate constraints into a search process, to relax constraints
when a conflict occurs and to diagnose poor solutions to the scheduling problem (Rodammer and
White, 1988).
13

Finally, the impractical assumptions are the main reason causing the problem in current
scheduling function. Those assumptions are listed by Larsen and Atling (1992) as follow:
A production plan (schedule) is assumed to be deterministic for the scheduled period.
This is only true when it is assumed that the manufacturing processes are stable, simple
and well understood with predictable and reliable setup and processing times. Likewise it
is assumed that the demand is predictable and reliable. By nature, manufacturing is non-
deterministic.
Scheduling follows the process planning in a strictly sequential order and it is restricted
to fixed process plans which are not to be altered.
Information on processing requirements and the resource status are assumed to be
accurate and complete.

2.3 Integration of Process Planning and Scheduling Function
Traditionally, process planning and scheduling for parts were carried out in a sequential way,
where scheduling was done after process plans had been generated. This sequential approach
ignores the inherent relationship between process planning and scheduling and process planners
and makes plans without considering the condition of resources on the shop floor, presuming an
idle shop floor (Kempenaers et al, 1996; Zhang et al, 1993). This is similar to Dong et al (1992)
research which identified that separating process planning and scheduling is resulting in the lack
of flexibility and adaptability in manufacturing.
Considering the fact that the two functions are usually complementary, it is necessary to
integrate them more tightly so that performance of a manufacturing system can be improved
greatly (Shao et al, 2008). The idea of this integration is driven by the need to increase the
flexibility and responsiveness to the shop floor (Usher and Fernandes, 1996). It is possible to say
that improvement of the interaction and integration between process planning and scheduling
could lead to a dynamic manufacturing environment. The integration is capable of reacting to a
14

real-time shop floor conditions, and can significantly reducing the cost and lead-time as well as
improving manufacturing responsiveness.

2.3.1 Approaches of the Integration
Over the last two decades, there have been numerous research efforts towards the integration of
process planning and scheduling. One of the very first approaches, invented by Chryssolouris
and Chang (1985), is manufacturing decision making (MADEMA). This approach identifies the
assignment of various factory resources to the production tasks as the common aspects of process
planning and scheduling functions. The integration problem is modeled as a multiple attribute
decision-making problem. A decision matrix is created where the rows represent alternatives, in
terms of resources while the columns represent attributes. The choice of one or the other
alternative resource is made by the evaluation of its relevant contributions to some established
criterion. Afterward, a number of different approaches of integration between process planning
and scheduling have been addressed by many authors.
Sundaram and Fu (1988) proposed an integrating methodology that stresses the assignment of
operations to alternative machines. This methodology seeks not only to minimize the makespan
but also to balance the load of machines.
Srihari and Greene (1988) discussed and described a CAPP system that consider the concept of
alternative routing in order to improve route efficiency, avoid bottlenecks, reduce in-process
inventory, level machine utilization, reduce flow time and streamline workflow.
Iwata and Fukuda (1989) proposed integration approached called Dynamic Process Planning
which decides on the process planning and scheduling functions without the customary prepared
alternative process plans. It sorts the input data for the system to include factors, for example, the
utilization status on the shop floor, the due date, the production volumes and the product design.
Khoshnevis and Chen (1990) introduced DYNACAPP (Dynamic Computer-aided Process
Planning), which combines process planning and scheduling functions. In the integrated system,
less costly schedules can be generated based on alternative process plans provided by the process
15

planning function. A priority dispatching method with concurrent assignment algorithm is
developed; it uses a time window scheme that controls the number of assignments made at each
stage. This integrated system provides a larger problem search space that creates more flexibility
in problem solving. To conclude, the mean of the processing time is the major factor that should
be considered when selection the time window, however, what the exact window size for a
specific problem should be needs further investigation.
Iwata et al (1991) described a structure of knowledge for an integration process planning system
(IPPS) which consists of process planning of machine process and production scheduling in the
job shops.
Dong et al (1992) worked to integrated process planning and scheduling by using CAPP system
which provided alternative machines to carry out the processes. The CAPP system considers
detachable volume shape for process selection but does not considered dimensional and
geometric tolerances. In addition, scheduling of task the system also does not take into account
set up times and material handling times. This would be able to move parts of a new machine,
which had the same capability as the previous.
Carvalho (1996) introduced the capability units termed resource elements (RE), which are
groups of form generating units (schemas) that always appear together in a machine tool. This
approach is based on a methodology of representing both the capabilities of a machining facility
and the requirements to produce components. Although these capability units are the basis of the
proposed system, they are applied in different ways, or as this work suggest, stages of
development. The remarkable advantage of RE based scheduling is that the information about all
possible ways of producing a component is used during the simulation.
Tan and Khoshnevis (1997) explained a linear mixed integer programming model (LMIP), which
integrates process planning and scheduling into one single formulation. The LMIP model can
accommodate a variety of objective functions for both process planning and scheduling needs,
for examples, minimizing makespan, minimizing production cost, maximizing on-time
performance, and minimizing the number of tardy jobs. In some cases, it does require longer
running time to obtain better solution quality.
16

Ferreira and Wysk (2001) illustrated that the routing alternatives can increase the efficiency of
manufacturing the resource such as increasing the resource utilization and reducing processing
time. A single machine is based on their case study and they assume an ideal situation that there
is no machine failures and maintenance.
Lim and Zhang (2003) introduced a multi-agent based framework in the integration of process
planning and scheduling problem. This framework can be used to optimize the utilization of
manufacturing resources dynamically as well as provide a platform on which alternative
configurations of manufacturing systems can be assessed. In fact, a typical definition of an agent
is given by (Nwana and Ndumu, 1997) as:
An agent is defined as referring to a component of software and/or hardware which is capable
of acting exactly in order to accomplish tasks on behalf of its user

Kim et al (2003) proposed a method for integrating process planning and scheduling in job shop
flexible manufacturing system. His work based on an artificial intelligent technique and also
concerns with sequencing and processing flexibility.
Wong et al (2006) described an online hybrid agent-based negotiation multi-agent system (MAS)
for integrating process planning with scheduling/rescheduling. With the introduction of the
supervisory control into the decentralized negotiations, this approach is able to provide solutions
with a better global performance.
Zhang and Xie (2007) developed the newly MAS which is more effective to solve the integration
of process planning and scheduling problem than single-agent approach. However, when the
number of the agents is large, agents will spend more time processing message than doing actual
work, and it is often difficult to apply the genetic agent architectures directly to integration
process planning and scheduling systems. Moreover, they indicated that a concept of an agent
came from the research of artificial intelligence (AI).
17

Shao et al (2008) proposed the algorithm-based approach for the integration of process planning
and scheduling (IPPS) in which synthesizes the methodology of NLPP and simulation. The basic
steps of this approach are as follows:
1) Process planning system is used to generate the alternative process plans for all jobs and
select user-defined number optimal plans based on the simulation results.
2) The algorithm in the scheduling system is used to simulate scheduling plans base on the
alternative process plans for all jobs.
3) Based on the simulation results, the process plan of each job and the scheduling plan are
determined.
Although this is a practical approach, the main disadvantage is that the simulation time may be
long and cannot be used in the actual manufacturing system.

2.3.2 Classification of the Integration Approaches
As aforementioned, there are many authors have developed a variety of approaches to
accomplish the integration of process planning and scheduling. In this part, several researchers
have classified them into different categories as follow:
1) Schmidt and Kreutzfeldt (1992) make the following division:
Dynamic process planning: This approach combines process planning and scheduling
into one single optimization task. This approach seems unfeasible for job shop
production.
Opportunistic planning: This approach does not consider all shop orders simultaneously,
but still generates process plans based on the actual shop situation (Khoshnevis, 1990).
This approach still allows a considerable time span between process plan generation and
the completion of the last operation on the shop floor. Replanning is not supported by this
approach.
18

Two level process planning: The operations are only roughly determined in the first step.
The detailed planning is performed at a much later point shortly before manufacturing.
Non-linear process planning: Scheduling and process planning still remain distinctive
functions with respect to time. Only one process planning is created for each component
in a Petri Net format. FLEXPLAN is based on this concept (Tonshoff et al, 1992).
2) Larsen and Alting (1992) and Shao et al (2008) have classified the proposed model of the
integration of process planning and scheduling into three types:
Nonlinear process planning (NLPP): The methodology of NLPP is to make all alternative
plans for each part with a rank according to process planning optimization criteria
(Beckendorff et al, 1991). For example, the highest priority plan is always submitted to
the job otherwise the second-priority plan will be given to the scheduling. However,
through a number of experimental computations, Usher (2003) concluded that the
advantage gained by increasing the number of alternative process plans for a scheduling
system diminishes rapidly when the number of the plans reaches a certain level.
FLEXPLAN is one example of such a system (Tonshoff et al, 1989).
Closed loop process planning (CLPP): It is using a dynamic process planning system
with a feedback mechanism from the shop-floor (Khoshnevis and Chen, 1989). The
process planning mechanism creates process plans based on available resources at that
time. Production scheduling tells process planning which machines are available on the
shop floor for an incoming job, so that every plan is feasible and respects to the current
availability of production facilities. Also, this dynamic simulation system can enhance
the real-time, intuition and manipulability of process planning system and enhance the
utilization of alternative process plans. Example systems include RTCAPP (Khoshnevis
and Chen, 1990) and DPP (Tonshoff et at, 1989).
Distributed process planning (DPP): The method involves performing both the process
planning and production scheduling in parallel with a hierarchical approach (Iwata and
Fukuda, 1989). It divides the process planning and production scheduling tasks into two
phases. The first phase is an initial phase. In this phase, the characteristics of parts and
19

the relationship between the parts are analyzed and the primary process plans are
determined. The process resources are also evaluated simultaneously. The second phase
is a detailed planning phase. In this phase, the process plans are adjusted to the current
status of shop floor. The detailed process plans and scheduling plans are obtained
simultaneously. IPPM (Zhang, 1993) and IPPS (Huang et al, 1995) are the examples.
Note that: A comparison table among these integration models is shown in Appendix A
3) Zhang et al (2003) classified the report approaches into two categories:
Iterative approach: under this category, the CAPP system and the scheduling system are
kept as two separate functional modules. For a given set of jobs, multiple feasible process
plans are generated for each job. A top prioritized plan for each job is then chosen and
input to the scheduling system for generating a schedule. If the generated schedule is not
satisfactory, a job is chosen and its current plan replaced by another alternative plan. This
iterative process continues until a satisfactory schedule is found or no further
improvement can be made. A typical example of this approach is the nonlinear process
planning (NLPP) approach (i.e. Tonshoff et al, 1989). The limitation among the reported
developed systems is the lack of intelligent search strategy for choosing an appropriate
process plan, so it makes them rather like a trial-and-error process.
Simultaneous approach: based on the idea of finding a solution (process plans for all the
jobs and a schedule) from the combined solution space of process planning and
scheduling. The basic elements are features that form the parts in given jobs. The
objective is to find a process plan for each feature and a sequence in which features pass
between machines subject to the technological constraints and some optimization criteria
with respect to process planning and scheduling performance. An obvious example of
this approach war provided by Tan and Khoshnevis (2000). The strength of this approach
is that the integration problem modeled in a truly integrated manner with the whole
solution space available. However, with such a vast solution space, finding a feasible
solution in a reasonable amount of time is too difficult.

20

2.4 Computer Simulation
Kelton et al (2003) defined simulation as a method used to create a model with the characteristics
of a real system on a computer with the appropriate software. Simulation is a powerful problem-
solving technique that is concerned with statistical sampling theory and analysis of complex and
probabilistic physical systems (Kelton et al, 2003). Currently, simulation studies have been
carried out in most business factors including manufacturing and service industries as well as in
the public sector (Robinson, 1994). In particular, application of simulation methods have been
found in production systems including material handling, process planning, inventory control
systems, production lines and job shop scheduling. Moreover, computer simulation is used for
solving engineering problems by experimenting on a computer based model (Robinson, 1994).
Its application has led to improvements in efficiency, reduces costs and increases profitability for
business, including those in the manufacturing and service industries.
From the practical viewpoint, simulation is the process of designing and creating a computerized
model of a real or proposed system for the purpose of conducting numerical experiments to give
us better understanding of the behavior of that system for a given set of conditions (Kelton et al,
2003). Carvalho (1996) suggested that although it is difficult to represent the activities of the
system for realistic models because of the complexity of the manufacturing system, simulation
models can be evaluated numerically and give the necessary realism. As well as the work of
Song et al (2006) which argued that simulation models can represent real-world systems at
almost any level of detail in order to approximate the actual system. Moreover, Pidd (1992)
indicated that the model which produces the best results would be implemented in the real
system (Pidd, 1992).
Scheduling simulation is modeled to provide the user with the capability of performing what if
analysis on the scheduling problem. This is supported by the work of Pidd (1992) as he noted
that computer simulation involves experimentation on a computer-based model of some system.
The model is used as a vehicle for experimentation, often in a trial and error way to
demonstrate the likely effects of various policies. The basic idea of this basic mechanism is
provided in Figure 2.1
21

Figure 2.1: Simulation as experimentation <Pidd, 1992>

Over the last two or three decades, simulation has been consistently reported as the most popular
operations research tool. The reason for this is its ability to deal with very complicated models of
correspondingly complicated system (Kelton et al, 2003). Another reason is the obvious
improvement in performance/price ratios of computer hardware, making it ever more cost
effective (Kelton et al, 2003). It is supposed that the reputation and effectiveness of the computer
simulation are now becoming even greater than ever due to the advance in computer hardware
and software.

Verification and Validation
Once the simulation model has been completed, it is time to make sure that the model is
performing as designed. Pidd (1992) stated that one important aspect of simulation model is the
models should be thoroughly tested or validated before use. The accuracy of the model is very
important in order to carry out the purpose of the project. Verification and validation are actually
important quality assurance procedures used in a simulation study. To start with, the objective of
verification is to certify the accurate performance of each element in the model (Robinson,
1994). It is important to repeat that verification of performance while the model is being
22

assigned. As the logic and data are being input into the computer model, it is necessary to ensure
that each element in the model is verified.
Apart from verification, the purpose of validation is to guarantee the correct degree of accuracy
by checking that the overall behavior of the model is representative of the real world (Robinson,
1994). Model validation is concerned with both accuracy and the purpose of the simulation
project for which it is being used. In addition, validation is not only to certify the correct degree
of accuracy by verifying it but that the overall performance of the model is representative of the
real world. In sum, verification and validation are different elements of simulation that both
ensure the accuracy and relevance of the simulation activity and its solution. Although
verification and validation are described as two separate tasks here, there is a significant amount
of overlap between them (Robinson, 1994).

2.5 Summary of the Research
In real-time manufacturing, an unexpected event, such as machine breakdown, lack of resource
availability and etc, can emerge during shop floor production inevitably. As process planning
requires ability to dynamically handle all those situations, the integration of process planning and
scheduling has to be prepared. In this case, the newly computer simulation method has been
applied in order to generate the dynamic planning regarding to the integration of those two
functions. The version 8.01 of Rockwell Software Arena has been used to store the shop floor
data and test the simulation model.
To begin the experiment, a set of machines and jobs along with processing times in each
operation, obtained from real company, are exploited to build a simulation model. Even if Arena
only replicates the situation and apply the data in the computer, it can effectively represent the
manufacturing system in the real-time situation. It also can investigate the characteristics of
manufacturing performance in different situations. By using Arena, the model is mainly built in
three aspects which consist of two static models and one dynamic model. The first static model is
referring to the method from previous project (Wongthanasunthorn, 2007). It has been built in
23

static manner by using the sequence method. However, the second-built static model has been
created in another way which is the variable method.
For the last model, it is constructed dynamically so as to cope with the unstable situation occurs
within the shop floor. In this case, the approach of alternative machine has been employed to
observe the effects in the shop floor level. The manufacturing performance is specifically
focused in terms of minimizing processing time, waiting time, number of waiting in queue
whereas increasing the productivity. In fact, four different simulation experiments will be tested
in this dissertation.

24

Chapter 3
Research Methodology

This chapter starts with the introduction of resource data using in the research. It is actually
based on the real shop-floor data from the company in Thailand. Then, it is going to describe the
steps necessary to build a simulation model and its subsequent functions by using Rockwell
Software Arena. Next, the simulation model section will be separated into three main parts
including static model (sequence method), static model (variable method) and dynamic model.

3.1 Data for Simulation Experiment
According to the previous project, Wongthanasunthorn (2007) acquired a real job-shop data from
Supreme Precision Manufacturing (SPM) enterprise in order to carry out the project. In this
research, the investigation also focused on the actual shop floor records from the same company
so as to test the simulation experiment. SPM, Thailand-based Company, produces components of
high precision with machining potentials to supply a number of factories worldwide. Since SPM
is widely known as the modern factory with high-technology equipment, the shop floor
department can be divided into two major areas which consist of CNC (Computer Numerical
Control) high precision technology and manual machining technology.
In reality, this significant set of data came from one particular production line of their shop floor
section. The data from the SPM mainly includes the production of five components which are
processed through seven different workstations. The high precision products contains flexible pin
base, bearing mount, motor stand, slide bush and wheel holder. For the workstations, used to
manufacture the parts, it comprises of cutting machine, lathe machine, milling machine, milling
face machine, CNC lathe machine, CNC milling machine and tapping machine. An amount of
25

ordered parts is presented in table 3.1 and the sequence of each part along with processing times
in each workstation is also provided in table 3.2

Item Description Quantity/Batch (pcs.) Material
1 Flexible Pin Base 100 SUS304
2 Bearing Mount 100 SUS304
3 Motor Stand 40 A6061
4 Slide Bush 50 SUS304
5 Wheel Holder 50 A6061

Table 3.1: The description of high precision parts that are produced by SPM enterprise including
the number of parts required. <Wongthanasunthorn, 2007>

Item
Quantity/Batch
(pcs.)
Cutting
(min)
Lathe
(min)
Milling
(min)
Milling Face
(min)
CNC Lathe
(min)
CNC Milling
(min)
Tapping
(min)
P1 P2 P1 P2 P1 P2 P3 P1 P2 P1 P2 P3
1 100 SQ1 SQ2 SQ6 SQ3 SQ4 SQ5
1 5 12 4 3 6
2 15 5 2 3 2
3 40 SQ1 SQ2 SQ3 SQ4 SQ5 SQ6 SQ7 SQ8
3 30 20 15 20 15 20 10
3.5 10 3 4 8 7
8 8 5 15 10 8

Table 3.2: The table represents the sequence for each item processing through the entire system
along with processing times in minutes. <Wongthanasunthorn, 2007>

26

3.2 Simulation Model Building
This section basically explains the key procedure to build the simulation model in Arena. The
models are proposed to investigate the integration of process planning and scheduling function
using actual shop floor data from one production line of SPM enterprise. There are several
assumptions to be made in the models building processes: 1) The areas of the factory, position of
human and mechanism of machines are neglected. 2) Transfer times and distances between each
machine are neglected. 3) Set-up times and delay times in each machine are neglected. These
assumptions are applied to simplify the complication of the simulation model structure.

3.2.1 Static Model using Sequence Method
According to the data provided in table 3.2, there are five part types that follow different routes
through the system. In the static condition, an item is processing through the system by following
just one specific route. Thus, an Arena concept called sequence, which easily allows the flow of
entities through the system, has been applied. Arena sends the entities through a system
automatically according to a predefined sequence of station visitations.
To start building model, Sequence data module from the Advance Transfer panel had to be
edited first. The Sequence data module let us define an ordered list of Stations that can include
assignments of attributes or variables at each station (Kelton et al, 2003). Thus, it is essential to
define and name the list of stations to be visited for each type of entity using the Sequence data
module. Figure 3.1 shows the procedure for Sequence of Item 1 Process Plan, Step of Item 1
Step 1 and Assignment of Process Time. Using the data from table 3.2, Item 1 Process Plan had
got 7 rows in the steps which meant that Item 1 (i.e. flexible pin base) was going through seven
stations including exit system in order to complete the manufacturing process. Importantly, the
most common error in entering the sequences is to forget to enter the last step which is typically
where the entity exits the system. For the attribute, processing time, it represents that the cutting
machine required 100 minutes (i.e. 100 pieces/batch*1 minute) to processing the job. An
attribute is basically a common characteristic of all entities, but with a specific value that can
differ from one entity to another (Kelton et al, 2003). To complete all Sequence data modules, it
27

is fairly similar to this example when setting up the Sequence data module for the item 2, 3, 4
and 5.

Figure 3.1: The Sequence data module of Item 1 Process Plan

Next, the main portion of the models operation consists of logic modules to represent item
arrivals, item processing and item departures. The item arrivals have been modeled using the
modules shown in figure 3.2. The detailed explanation can be given by this example; the Create
module, Item 1, used a Constant distribution with a value of 40 hour to generate the arriving
items (see Appendix B, figure B2).

28

Figure 3.2: Logic of the item arrival modules for producing item 1

For Assign modules, they have been used to define two assignments as shown in figure 3.3. An
index, Part Index, was generated as attribute. It is essential to define the value of Part Index first
in the Assign module, since its value would be used in the later assignments. Also, they assigned
the proper sequence to the Arena attribute, Entity.Sequence, by using the Part Index attribute as
an index into the Part Sequence set.

Figure 3.3: A detailed of Assign module, Assign Item, in item arrivals chain. The example
assignments of item 1 are also represented as it assigns Part Index attribute and Entity.Sequence
attribute.

29

After that, it was prompt to send the parts on their way to the first station in the sequence. Leave
module from the Advance Transfer panel directed the entities to follow the pattern of station
visitations. This Start Sequence module has been set to connect type as Route and connected the
station by Sequence. This caused Arena to route the arriving entities according to the defined
sequence and attached to entities after they were created. The detailed of Start Sequence is
provided in figure 3.4.

Figure 3.4: A detailed of Leave module, Start Sequence.

Now that the arriving items were routed according to their assigned part sequence, it now needed
to develop the logic for item processing. The logic is that an item arrives to the station, queues
for a machine, be processed by the machine and be sent to its next step in the item sequence. The
example of the item processing of Cutting operation is shown in figure 3.5.
30

Figure 3.5: Logic of item processing modules from Cutting machine. It includes Enter, Process
and Leave module.

The Station module provides the location to which a part can be sent. In this model, all items are
transferred to the next station from their defined sequence. Moreover, the processing time in the
Process module has been set as Expression of Processing Time which has already assigned in the
Sequence data module. Finally, having completely defined all the data, the items exit was the
only thing left to define. The two modules used to accomplish this are shown in figure 3.6.
Furthermore, the completed static model using sequence method is provided in Appendix B,
figure B1.

Figure 3.6: Logic of item departures for the whole system.

31

3.2.2 Static Model Using Variable Method
In the previous model, the simulation process has been executed by sending the items through a
system automatically based on the predefined Sequence data module. However, another
approach, using Variable data module, for building static model has been introduced in this
section. A variable is a piece of information that reflects some characteristic of your system,
regardless of how many or what kinds of entities might be around (Kelton et al, 2003). In
contrast to attributes, variables are not tied to any specific entity, but rather pertain to the system
at large.
The first step of model construction was to set the values in the Variable data module from Basic
Process panel. According to figure 3.7, the variables of Machines and Optimes were generated in
the data module. Machines variable defined the production routes in which all items need to take
while Optimes variable described the processing time in each machine on the routes. With the
data from table 3.2, they were placed in the module as shown in figure 3.7. Inside Initial Values
window, the vertical axis presented a type of item, like item 1 representing the flexible pin base
whereas the horizontal axis presented the sequence of the operation. From the example in figure
3.7, it shows the initial values for Machines variable which had 5 rows and 9 columns. Apart
from this, Optimes variable would clearly be completed in a similar way (See Appendix C, figure
C2).

32

Figure 3.7: The Variable data module of the static model using variable method. It includes the
detailed of initial values applied to Machines variable.

Next, the logic of item arrivals has been changed by adding one Submodel between the Item
Order Release Station and Start Process (see figure 3.8). The reason for adding this Submodel is
to assign the correct machines name to the numbers which have already been set in the Variable
data module. Referring to the figure 3.7, the names of machine have been replaced by a number
due to the software problem.

Figure 3.8: The logic of item arrivals modules for static model using variable method.

33

For Assign modules, they have been used to define four assignments as shown in figure 3.9. An
attribute of NextMachine and Process Time were assigned to interact with the previously defined
routes in the Variable data module.

Figure 3.9: The detailed of assignments of the Assign Item 1 module.

To send the items to their first station, Leave module was set up by a new method (see figure
3.10). The Start Process module has been set to connect type as Route but this time it connected
the station by Attribute named Route To. This let Arena to route the arriving entities according to
the defined routes set in the Variable data module.
34

Figure 3.10: The set up details of Leave module in order to route the items according to the
defined Variables.

Considering the logic for item processing (see figure 3.11), the additional Assign module and
Submodel were employed. These further modules were used to send the items to the next
workstations after they finished this process. Finally, the overall model of the static model using
variable method is represented in the Appendix C, figure C1.

Figure 3.11: The logic for item processing when the item goes through the Cutting process.
35

3.2.3 Dynamic Model Using Variable Method
Unlike the static process planning, this planning has to be generated dynamically in order to deal
with unforeseeable event in the shop floor. In this part, dynamic simulation model will be built
together with the consideration of alternative sequence. Alternative machine concept is
specifically applied for creating this model. A list of the alternative sequences, machines and
processing time are provided in the Appendix E.
As this dynamic model uses the same variable method as the model presented in section 3.2.2,
the structure of the model is quite similar except some additional modules. Firstly, the Variable
data module needed to be set. According to figure 3.12, the variables of Machines, Optimes and
QLS were generated in the data module. The QLS was set to define an availability of the queue in
each machine. It expected to be used later in the decision process. As you can see from figure
3.12, the row of the Initial values was increasing to 10 rows. The values in 6
th
row to 10
th
row
represented the availability of the alternative machines. Apart from this, Optimes and QLS
variables will clearly be filled in a similar way (See Appendix D, figure D2 and D3).

Figure 3.12: The Variable data module of the dynamic simulation model. It consists of the
detailed of initial values applied to Machines variable.
36

Then, the logic of item arrivals has been modified by adding another Submodel locating after
Item Order Release Station (see figure 3.13, Submodel 18). The reason for adding this Submodel
is to assign logic of choosing alternative machines to the producing parts. The detailed of this
logic is provided in figure 3.14. The item will decide whether it has a choice of machines or not
(see detailed in Appendix D, figure D4). If there is no alternated machine, it will go to the next
machine according to the defined Variable data otherwise it will go to the next Decide module.
The second Decide module indicates that the item will choose the machine which has a lower
queue as its next destination (see detailed in Appendix, figure D5).

Figure 3.13: The logic of item arrivals of the dynamic model which contains two Submodels.
This is the example from the item 1 process.

Figure 3.14: The logic inside the new added Submodel in the dynamic model. This represents
that the item will decide which machine is more preferred to go.

37

The detailed of Assign modules and Leave module was set up exactly in the same way as the
static model using variable method. The items would be routed via attribute named Route To
which was scheduled them regarding to the defined Variable data module. Furthermore, the logic
for item processing (see figure 3.15), the Submodel module has been employed in the purpose of
deciding the suitable route for the items. At last, the overall model of the dynamic model using
alternative machine concept is shown in the Appendix D, figure D1.

Figure 3.15: The item processing part in the dynamic model dealing with alternative machine
concept. The Cutting process is used as an example.

38

Chapter 4
Experiments and Results Analysis

In this research, the simulation experiment is used to study the integration of process planning
and scheduling in the shop floor level. The investigation begins with the creation of a model
frame by the use of Arena software. As presented in the last chapter, the simulation models are
built in three different structures which exhibit the different process planning implementing in
manufacturing system. They also reveal the characteristics of the manufacturing performance in
dissimilar circumstances. The first model has been built in the static manner as well as the
second model but using different methods. For the third model, it was constructed for the
intention of creating the dynamic process planning. By using the concept of alternative
machines, the dynamic shop floor condition was generated.
Inside this shop floor, it contains seven machines for dealing with the production of five different
high precision products. The list of these machines and their capabilities is provided in table 3.2.
The objective of this chapter is to provide and analyze the results of experiments prepared with
those models. The performance indicators are mainly considered in terms of the amount of item
coming out, production time, time waiting in queue, number of item waiting in queue and utility
of the machine.

4.1 Experiment 1: Comparison of Static Model using Sequence Method and
Static Model using Variable Method
The first experiment begins with the evaluation of two models which similarly represent the
static process planning in the shop floor manufacturing system. The mechanism of the first
model is set to replicate the manufacturing system by using sequence module approach. This
data module is used to define an ordered list of Stations to be visited along with the processing
time in each operation. However, the second model is built statically with the use of variable
39

module which can also send the item through the system according to the defined variable. The
main purpose of this experiment is to check whether the models are acceptable built or not. In
this case, a highly 1,000 hours of replication length was set in the virtual manufacturing shop
floor so as to obtain the clearly practical results. After running the simulation for one replication
of 1,000 hours in these two models, the results of these simulations are compared in table 4.1 and
table 4.2.

Static (sequence) Static (variable)
Item No. out (pcs.) Time (hr) No. out (pcs.) Time (hr)
1 12 51.67 12 51.67
2 12 48.33 12 48.33
3 6 88.67 6 88.67
4 8 29.58 8 29.58
5 14 45 14 45
Total 52 52

Table 4.1: Simulation results of Static model using sequence method and Static model using
variable method. It shows the number produced items and production time for each item.

40

Static (sequence) Static (variable)
M/C
Waiting time
(hr)
No. waiting
(hr)
Utilization
Waiting time
(hr)
No. waiting
(hr)
Utilization
Cutting 3.5545 0.359 0.2479 3.5545 0.359 0.2479
Lathe 185.92 18.7131 0.9983 185.92 18.7131 0.9983
Milling 0.8333 0.01 0.24 0.8333 0.01 0.24
Milling face 0 0 0.4375 0 0 0.4375
CNC lathe 3.5367 0.4533 0.8308 3.5367 0.4533 0.8308
CNC Milling 120.11 16.615 0.9758 120.11 16.615 0.9758
Tapping 0.2604 0.008 0.1733 0.2604 0.008 0.1733
Total 314.2149 314.2149

Table 4.2: Table representing the results of two static models which include the comparison of
waiting time and number of waiting in each process as well as machines utilization.

As you can see from the table 4.1, the amounts of produced item and the production time of the
two static models were obviously the same. Flexible pin base (item 1) and bearing mount (item
2) alike were manufactured in 12 pieces. While 6 pieces of motor stand (item 3) were produced,
slide bush has been produced by 8 pieces. The final item (i.e. wheel holder) was built as the
largest amount which was 14 pieces. For the production time, item 3 was the one using the
longest average time to complete (i.e. 88.67 hours) while item 4 was the fastest item to be
produced which used just only 29.58 hours to manufacture. Item 1, 2 and 5 were typically built
in 51.67, 48.33 and 45 hours respectively.
The results were also similar to the outcomes in the table 4.2. Waiting time, number of waiting
and machines utilization of every machine were apparently identical for both static models. The
longest of average waiting time in the queue was 185.92 hours which occurred during the lathe
machine operation. Also, the remarkable long waiting time was happened in the operation of
CNC milling which was 120.11 hours. Moreover, the items have spent less than 4 hours in the
queue for the rest of the machines particularly the milling face machine which was no waiting
time at all. The overall waiting time in this static condition was 314.2149 hours.
41

For number of the item waiting in the queue, a large number of items were lined up for the lathe
machine process and CNC milling process. There were 18.7131 items and 16.615 items waiting
for processing to the lathe machine and CNC milling machine. Finally, the results of machines
utilization were clearly high in lathe machine, CNC lathe machine and CNC milling machine.
There are 0.9983, 0.8308 and 0.9158 consequently. Furthermore, the lowest utilization of
machine was during the process of tapping machine which was 0.1733. Overall, it is possible to
say that these two static models were proved reasonably built regarding to those identical results
shown in table 4.1 and 4.2.

4.2 Experiment 2: Comparison of Static Model and Dynamic model
This experiment is created to provide evidence that process planning which includes the
alternative routes can outperform the traditional one fixed process planning. It is expected that an
availability of alternative paths is critical decision so as to improve the performance of a
manufacturing system. Particularly, the alternative machines approach has been applied as a
basis of constructing this dynamic simulation model. The list of alternative machines with
processing time for particular operation is provided in the Appendix E. To inspect the difference
between the dynamic and static environments, the simulation was run for one replication of
1,000 hours. Consequently, the results of the simulation are compared as shown in table 4.3 and
4.4.

42

Static model Dynamic model
Item No. out (pcs.) Time (hr) No. out (pcs.) Time (hr)
1 12 51.67 18 22.037
2 12 48.33 20 23.33
3 6 88.67 15 75.5556
4 8 29.58 20 13.9167
5 14 45 24 44.0278
Total 52 97

Table 4.3: The simulation results of Static model and dynamic model. It compares between the
number produced items and production time for each item.

Static model Dynamic model
M/C
Waiting time
(hr)
No. waiting
(hr)
Utilization
Waiting time
(hr)
No. waiting
(hr)
Utilization
Cutting 3.5545 0.359 0.2479 3.4476 0.3344 0.2413
Lathe 185.92 18.7131 0.9983 2.7155 0.3476 0.6029
Milling 0.8333 0.01 0.24 2.5833 0.052 0.3458
Milling face 0 0 0.4375 0 0 0.4375
CNC lathe 3.5367 0.4533 0.8308 105.28 12.8372 0.9917
CNC Milling 120.11 16.615 0.9758 8.0556 0.7733 0.9317
Tapping 0.2604 0.008 0.1733 0.2083 0.005 0.16
Total 314.2149 122.2903

Table 4.4: Table representing the results of static model against dynamic model which include
the comparison of waiting time and number of waiting in each process as well as machines
utilization.

43

According to the table 4.3, the number of produced item within dynamic situation was
significantly increasing comparing to static planning system. The amounts of item 3 and 4 were
rising approximately twofold. Item 3 and 4 were produced up to 15 items and 20 items instead of
6 items and 8 items in the static model. Then, the largest number of produced item was item 5
which were built in 24 pieces rather than previously 14 pieces. Actually, the overall item being
produced from the dynamic model was 97 items which was higher than 52 items from the static
model. For the production time, the results gathered from dynamic model were notably
decreasing. Especially, the average time required to finish item 1 and 4, they were spending
around half of the times used in the static condition. Within the dynamic condition, item 1, 2, 3,
4 and 5 were spending 22.037, 23.33, 75.56, 13.92 and 44.03 hours consequently in order to
manufacture one of the item.
Considering the results in table 4.4, it can be said that the waiting time of most machines in
dynamic model were falling by far. For instance, waiting time in the line for lathe machine was
reduced from 185.92 hours in static model to only 2.7155 hours in dynamic model whereas the
waiting time was dropped from 120.11 hours in static model to 8.0556 hours in dynamic model
according to the operation of CNC milling. Nevertheless, there were two machines which stood
in the line longer in the dynamic model. Milling machine dynamically spent about 2.5833 hours
when it used about 0.8333 hours in static condition. Similarly to CNC lathe machine which
wasted 105.28 hours waiting in the line in dynamic model when it used just 3.5367 hours in
static model. In fact, there was no item wasting time in queue for using milling face machine
either in dynamic or static model as the zero value shown in the table 4.4.
For the number waiting, the results showed a similar trend as presented in the waiting time. As
most machines had shorter waiting time in dynamic model, they had lower number of item
waiting in the queue as well. The biggest number of item waiting was in the process of CNC
lathe which was 105.28 items. Apart from this, the other machines in dynamic model had less
than 1 item waiting in the line for operations. For example, 0.334 and 0.3476 items have waited
for the cutting machine and lathe machine respectively in dynamic situation. Additionally, the
utilization of machine was quite high during the process of CNC machine and CNC milling
which were 0.9917 and 0.9317 accordingly in dynamic model rather than lathe machine and
44

CNC milling in static machines. The lowest utilization in dynamic model was the same operation
as in the static model which was tapping machine. In sum, it can possibly be said that the
performance of manufacturing system was improved when the dynamic process planning has
been achieved.

4.3 Experiment 3: Effects to Manufacturing System after Changing the
Replication Length
This experiment is prepared to inspect how the manufacturing performance changes when the
system is running for a much longer time. Due to the continuous run of machine in the actual
factory, the experiment is worthwhile trying. To start the experiment, the replication length of
dynamic model has been increased to four different values which were 1,000 hours for the
original value, 5,000 hours, 10,000 hours and finally 20,000 hours. For the information of
machine routes, alternative plans and processing time implementing in this experiment is given
in the Appendix E. As a result, the outcomes of the simulation with different replication lengths
are shown in table 4.5 and 4.6
Rep. length =1,000 hr Rep. length =5,000 hr Rep. length =10,000 hr Rep. length =20,000 hr
Item
No. out
(pcs.)
Time (hr)
No. out
(pcs.)
Time
(hr)
No. out
(pcs.)
Time
(hr)
No. out
(pcs.)
Time
(hr)
1 18 22.037 102 19.9183 184 24.68 381 26.3386
2 20 23.33 102 23.33 200 23.33 395 23.33
3 15 75.556 83 75.6064 161 75.63 313 75.52
4 20 13.9167 102 14.1176 200 14.14 395 14.15
5 24 44.0278 124 44.6237 249 44.17 499 44.185
Total 97 513 994 1983

Table 4.5: The simulation results of dynamic model when running in four different replication
lengths. It compares between the number produced items and production time for each item.
45

Runtime =1,000 hr Runtime =5,000 hr
M/C
Waiting time
(hr)
No. waiting
(hr)
Utilization
Waiting time
(hr)
No. waiting
(hr)
Utilization
Cutting 3.4476 0.3344 0.2413 3.3185 0.3206 0.2419
Lathe 2.7155 0.3476 0.6029 3.4073 0.4368 0.5996
Milling 2.5833 0.052 0.3458 2.9545 0.065 0.38
Milling face 0 0 0.4375 0 0 0.4375
CNC lathe 105.28 12.8372 0.9917 455.47 54.3654 0.9983
CNC Milling 8.0556 0.7733 0.9317 9.2133 0.8907 0.9582
Tapping 0.2083 0.005 0.16 0.3194 0.0076 0.16
Total 122.2903 474.683

Runtime =10,000 hr Runtime =20,000 hr
M/C
Waiting time
(hr)
No. waiting
(hr)
Utilization
Waiting time
(hr)
No. waiting
(hr)
Utilization
Cutting 3.3748 0.3243 0.2413 3.3238 0.3159 0.2396
Lathe 3.6416 0.4763 0.6214 3.568 0.4661 0.6203
Milling 2.822 0.062 0.383 3.35 0.075 0.392
Milling face 0 0 0.4375 0 0 0.4375
CNC lathe 952.27 118.49 0.9992 2024.57 254.19 0.9996
CNC Milling 9.1104 0.0901 0.9585 8.8365 0.8745 0.9598
Tapping 0.2909 0.0067 0.1547 0.2278 0.0052 0.1537
Total 971.5097 2043.8761

Table 4.6: The simulation results of dynamic model which testing in four different replication
lengths. They include the comparison of waiting time and number of waiting in each process as
well as machines utilization.

46

Referring to the table 4.5, it can be said that the number of item being produced within dynamic
situation was continuously increasing as the replication lengths rising. All of the items have been
produced by 97 items in 1000 hour runtime, and then it started increasing the amount of product
to 513 items, 994 items and 1,983 items in 5,000 hours, 10,000 hours and 20,000 hours of
replication length consequently. For the average time needed to build one piece of item, the trend
of results was practically stable. As you can see, item 1 was used around 20-26 hours to build the
item after testing all experiments. Apart from this, item 2, 3, 4 and 5 were spending
approximately 23, 75, 14 and 44 hours respectively to complete one piece of the product
regardless of the runtimes.
To consider the overall waiting time, the results tended to increase as the additional of replication
length applied (see table 4.6). Especially, the CNC lathe machine which the waiting time was
raising substantially as the time passed. It rose from 105.28 hours in the first test to 2,024.57
hours in the last test. Yet, there was only milling face machine that offered no queue at all. For
the number of waiting results, they showed quite a fluctuated trend. For instance, waiting time in
the line for lathe machine was increasing from 0.3476 items to 0.4763 items after running time
changed from 1,000 hours to 10,000 hours but it was reduced to 0.4661 items when the
replication length was set at 20,000 hours. Nevertheless, it is noticeable that the CNC lathe
machine have got significantly large amount of item waiting comparing to the rest machines.
Apart from this, the results of machine utilization were collected. There were two machine, CNC
lathe and CNC milling, which got relatively high utilization. They normally reached the 90%
usage in all tests. On the other hand, the low rate of utilization was in tapping machine and
cutting machine. The usage of these two machines was around 10-20%.

4.4 Experiment 4: Effects to the Manufacturing Performance as the Number
of Alternative Routes Increasing
To improve the dynamic process planning, the alternative paths have been considered vital. It is
also believed that the more number of alternative plans creating, the more development of the
process plan can be achieved. Thus, this experiment is prepared to analyze and explore the
47

influence the number of alternative routes creates. The examination started by running the
original dynamic model and then the extra number of alternative machine would be added
randomly to the next model. In the end, the last model, dynamic model C, would get the most
choices of machines to choose. The lists of alternative machines and their capabilities for all
dynamic models are presented in the Appendix F. After running one replication of 1,000 hours in
each model, the results are shown in table 4.7 and 4.8.

Original dynamic model Dynamic model A Dynamic model B Dynamic model C
Item
No. out
(pcs.)
Time (hr)
No. out
(pcs.)
Time (hr)
No. out
(pcs.)
Time (hr)
No. out
(pcs.)
Time
(hr)
1 18 22.037 24 10 23 7.5362 15 47.11
2 20 23.33 21 34.5238 23 36.4493 14 53.57
3 15 75.556 17 88.3137 22 86.4242 15 85.33
4 20 13.9167 21 26.9841 23 28.9855 13 30.3846
5 24 44.0278 24 45 24 46.3889 16 52.91
Total 97 107 115 73

Table 4.7: Simulation results of dynamic model with different number of alternative machines.

Original dynamic model Dynamic model A
M/C
Waiting time
(hr)
No. waiting
(hr)
Utilization
Waiting time
(hr)
No. waiting
(hr)
Utilization
Cutting 3.4476 0.3344 0.2413 6.0018 0.8523 0.7146
Lathe 2.7155 0.3476 0.6029 2.4845 0.2932 0.5163
Milling 2.5833 0.052 0.3458 1.2173 0.034 0.5108
Milling face 0 0 0.4375 0 0 0.4375
CNC lathe 105.28 12.8372 0.9917 62.0644 7.0138 0.9917
CNC Milling 8.0556 0.7733 0.9317 5.8546 0.55 0.9117
Tapping 0.2083 0.005 0.16 0.3627 0.0185 0.3
Total 122.2903 77.9853

48

Dynamic model B Dynamic model C
M/C
Waiting time
(hr)
No. waiting
(hr)
Utilization
Waiting time
(hr)
No. waiting
(hr)
Utilization
Cutting 6.2273 0.9222 0.7722 2.4289 0.2478 0.3573
Lathe 0.9621 0.1078 0.4946 102.17 10.1675 0.9575
Milling 5.3688 0.2657 0.8158 53.77 3.8043 0.9779
Milling face 1.9507 0.1912 0.6758 30.5536 3.7303 0.97
CNC lathe 14.9077 1.6673 0.9875 4.7946 0.4747 0.5542
CNC Milling 7.4364 0.7272 0.92 41.6304 4.9103 0.9608
Tapping 0.387 0.023 0.32 0.4267 0.021 0.1447
Total 37.24 235.7742

Table 4.8: The simulation results of dynamic model which testing in four different alternative
plans. The results include the comparison of waiting time and number of waiting in each process
as well as machines utilization.

In accordance with the table 4.7, as the number of alternative paths increasing in model A and B,
the amount of item being produced was clearly improving. The original dynamic model, the
fewest plans model, has built around 97 items while Dynamic model A and B, which had
relatively more plans, built about 107 and 115 items consequently. However, the dynamic model
C, which had largest number of plans, conversely produced the items in the lowest amount (i.e.
73 pieces). For the results of time used to make one piece of item, the trend was unstable. They
could either be raised up or fallen down when the number of alternative machines increasing.
Apparently, it can be said that item 3 required the longest time to manufacture among all the
items no matter how many alternative plans were.
Regarding to the table 4.8, the total average time spending in the queue was obviously growing
when the more choices of machines were employed in the model. They were decreasing from
122.29 hours to 77.99 hours and then 37.24 hours in dynamic model B. However, there was the
exception for the dynamic model C. Even if it had a largest number of plans, it stilled spending
time waiting in the line as long as 235.7742 hours. It is interesting that the tapping machine was
49

yet only machine that presented no queue in original model and model A but the queue was
presented after the simulation of model B and C. To focus on machine utilization, around 90 %
of the two CNC machines were used during the manufacturing processing in the original
dynamic model, model A and model B. In particular model C, there were four machines which
had fairly high rate of utilization. Lathe, milling, milling face and CNC milling were those
machines.

50

Chapter 5
Discussion

Once executing the simulation experiments, those obtained results will be critically assessed and
commented in this chapter. It is divided into four parts to mention each experiment.

5.1 Discussion of Experiment 1
The first experiment was generated to inspect the characteristics of two models which similarly
represent the static process planning in the shop floor manufacturing system. By using the
sequence method, the items have been sent all the way through the system in the first model.
Unlike the first one, the second model was constructed using variable data module to store the
detailed of routes. Considering the results in table 4.1, the amounts of produced items and the
production time of the two static models were exactly the same values. There were totally 52
items which produced in both static models. For the production time, it is obviously that even
item 3 was the one using the longest average time to complete (i.e. 88.67 hours), it produced just
only 6 items. Thus, it can be said that item 3 was the most difficult item to be built. Moreover,
the item 3 was the product which had the longest routes (i.e. had 8 operations) among all the
items according to the table in table 3.2 in chapter 3.
From table 4.2, waiting time, number of waiting and machines utilization of every machine
were identical for both static models as well. Lathe and CNC milling machine simply were the
two devices that spent considerable long waiting time and offered relatively longer queue
comparing to the other machines. It is possibly to say that these two machines could be the
favorite choices for this shop floor department. CNC milling machine, in particular, is the core
machine being used in the production of every item (see table 3.2). In fact, the overall waiting
time of this static system was 314.2149 hours. To consider the numbers of item waiting, a large
51

amount of items needed to wait for lathe and CNC milling machine processes while the other
machines queue were relatively small. They contained less than one item waiting in the queue
especially the milling face machine which was no queue at all. Consequently, this also caused
the superior rate of utilization in lathe and CNC milling machine.
Finally, as the intention of this test is to validate the simulation models, the experiment is
apparently proved that the model is built acceptably. As Kelton et al (2003) indicated that once
simulation model have been built completely, it is really important to ensure that the model is
performing as expected. Furthermore, at some point, it is almost impossible to verify totally a
model in a very complex system (Kelton et al, 2003)

The aim of this study is to compare results in terms of performance obtained when considering
either a single fixed route approach (i.e. static model) or an approach of alternative routes (i.e.
dynamic model). In the single fixed route approach, routes are created for each job always using
the same preferable machines. These routes must be followed by the production channel
personnel without any changes. On the other hand, alternative routes, specifically alternative
machines in this case, can generate the flexibility to the shop floor system.
In consideration to the table 4.3, the number of item being produced within dynamic situation
was dramatically increasing comparing to static planning system. The overall produced item
from dynamic planning was 97 items which was quite higher than 52 items from the static plan.
For the production time, the results gathered from dynamic model were obviously decreasing in
all items. They spent about 22.037, 23.33, 75.56, 13.92 and 44.03 hours in order to manufacture
one piece of item 1, 2, 3, 4 and 5 consequently. Hence, it can be said that the development of
manufacturing performance could be achieved by considering the process planning dynamically.
The reason for this development is that the workers do have choices to complete the items in the
more comfortable machine (i.e. smaller queue) rather than the busy one (see alternative routes in
the Appendix E). Additionally, the item 3 evidently required the longest time to manufacture
either in static or dynamic model.
52

In accordance with the results in table 4.4, most of the waiting time and number of item waiting
has been considerably reduced except in the process of milling and CNC machine. Milling
machine dynamically spent about 2.5833 hours when it used about 0.8333 hours in static
condition whereas the CNC lathe machine wasted 105.28 hours waiting in the line in dynamic
model rather than just 3.5367 hour waiting in static model. The reason could be that most of the
works have been transferred to these two machines according to the defined alternative plans.
Although, the waiting time has increased in the two machines, this caused a lot of decreased in
queue time for other machines. The overall waiting time was positively reduced from 314.21 to
122.29 hours which represents a pretty good improvement in process planning. For number of
item waiting in the line, there was only the CNC lathe machine in dynamic manner that had more
than one item staying in the queue. Moreover, the utilization of machine was quite high during
the process of CNC machine and CNC milling which were 0.9917 and 0.9317 in dynamic model
which was higher than lathe machine and CNC milling in static machines. It indicates that the
machines operating in the dynamic situation certainly perform better than the ones in static
situation.
Since all machines have their own utilization, it means that the machines are running correctly in
the simulation. Therefore, it can possibly be said that this dynamic model has been constructed
acceptably. Overall, it can be concluded that the performance of manufacturing system is
substantially improved when the dynamic process planning has been achieved. As suggested by
Sundaram and Fu (1988), if alternative machines are indicated when available, it will help avoid
overloading machines that are highly productive. Carvalho (1996) also argued that the use of
alternative routes improves the shop floors overall performance regardless of the efficiency of
the scheduling system in use. In fixed process plans, it often leads to schedules that end up with
severely unbalanced resource loading, create excessive bottleneck machines, and lead to lower
overall resource utilization and poor on-time delivery performance (Tan and Khoshnevis, 2000).

53

To inspect the manufacturing performance in more reality situation, the experiment was arranged
by increasing the replication length in the simulation program. The running period of dynamic
model has been increased to four different values which were 1,000 hours for the original value,
5,000 hours, 10,000 hours and finally 20,000 hours. Referring to the outcomes shown in table
4.5, the number of item being produced was steadily increasing as the replication lengths rising.
As expected the overall product was shifted from 97 items in 1,000 hour runtime to 513 items in
5,000 hour runtime, then 994 items in 10,000 runtime and finally 1,983 items in 20,000 hours
runtime. For the average time needed to build one piece of item, the trend of results was quite
steady. Item 1 was used around 20-26 hours to build while item 2, 3, 4 and 5 were spent
approximately 23, 75, 14 and 44 hours respectively to complete one product. Thus, this can
demonstrate the capability of the manufacturing production in this entire system.
To consider the results of waiting time in table 4.6, they might be varying in an individual
machine since the running period went longer, but the overall tended to increasing. The waiting
time of CNC lathe, in particular, rose from 105.28 hours in the first test to 2,024.57 hours in the
last test. For the number of waiting results, it is visible that the CNC lathe machine have got
considerably huge amount of item waiting comparing to the rest machines after the running time
constantly growing. The rest of the machines were produce no more than 1 item waiting actually.
This can be concluded that the CNC lathe is the most preferable machine within this shop floor
department. As you can see from the waiting time and number of waiting, this machine can
become the bottleneck which cause a serious trouble to the system sooner or later. Therefore, it is
necessary to prepare the plan to cope with this situation.
Usher and Fernandes (1996) supported the study of Carvalho (1996) that traditionally process
planning has been performed assuming infinite availability of resources. The production route is
normally created by selecting the machines that would perform the necessary operations better
which is the CNC lathe in this case. This may lead to the process planners favoring to select the
desirable machines repeatedly. Lee and Kim (2001) made it clear that these process plans are
somewhat unrealistic and cannot be readily executed on the shop floor. In reality, the
manufacturing resources are always insufficient and fixed routes can lead to bottlenecks, low
54

levels of machine utilization, long lead times, high levels of WIP, impossibility of meeting due
date and so on.

This experiment was organized to confirm that the more alternative routes approach can make
more improvement to the dynamic process planning as well. The inspection started by adding the
extra alternative routes to the dynamic model and run the simulation (see detailed in the
Appendix F). The original model had the lowest amount of plans and then increasing through
model A, B and eventually the highest number of plans was applied to model C. According to
the table 4.7, the amount of item being produced was relatively increasing as the number of
alternative paths increasing in model A and B. The original dynamic model, the fewest plans
model, had made 97 items while Dynamic model A and B, which had relatively more plans, built
about 107 and 115 items consequently. Conversely, the 73 items had produced in the model C
which got a load of plans more than model A and B. For the results of time required to make the
item, the trend was fluctuated.
Referring to the table 4.8, the trend of total waiting time was similar to the results of items being
produced. When the model A and B were simulated, the waiting times were dropping from
122.29 hours in original model to 77.99 hours and then 37.24 hours in model B. However, there
was the exception for the dynamic model C which spent time waiting in the line around
235.7742 hours. To consider the number of item waiting, CNC lathe was the only machine that
got the values more than 1. Interestingly, the model B was not only spent less time waiting, it
also had only 1.6673 items waiting for CNC lathe operation. The worst situation was took place
in the model C which got too many item waiting in the row. Milling, milling face and CNC
milling machine was presented 3.8043, 3.703 and 4.9103 items in queue respectively. The
biggest number of item in queue (i.e. 10.1675) was shown in the line for lathe machine. As a
result, these four machines had a reasonably high rate of utilization which was more than 90%.
For machine utilization in the first three models, just CNC lathe and CNC milling machines were
used around 90% during the shop floor production.
55

All in all, it can be said that the incremental of alternative plans can practically develop the
manufacturing performance. Carvalho (1996) illustrated that the number of routes considered is
very important for the expected performance of the system. The higher the numbers of the
routes, the better solutions are achieved and the more amount of computing time is required.
However, the appropriate number of plans needed to consider critically as well. According to the
literature of Larsen and Atling (1992), it described that there is an upper limit for how big an
addition in processing time can be when the alternative routings should be use with advantage. It
is because during use of alternative routings the alternative production resources often require
extra processing time to perform the same work as the first choice resource. Furthermore, J oo et
al (2001) indicated that considering all the possible routing alternatives for resource allocation
may enormously increase the complexity of process plan representation.

56

Chapter 6
Conclusion

The main purposes of this chapter are to outline the conclusion of the research project and to
provide the recommended further work in the area of the dynamic process planning regarding to
the integration of process planning and scheduling.

6.1 Research Conclusion
At the present time, global competition and rapidly changing customer requirements are forcing
major changes in the production styles and configuration of manufacturing organizations (Shen
et al, 2006). These complexities motivated industry to pursue good production schedules that
have great adaptability and flexibility (Chen and Chen, 2008). Manufacturing strategies should
shift to support global competitiveness, new product innovation and customization, and rapid
market responsiveness. The next generation manufacturing systems will thus be more strongly
time-oriented (or highly responsive), while still focusing on cost and quality (Shen et al, 2006).
Traditionally, process planning is executed without considering the dynamic behavior of shop
floor due to production scheduling, workload and traffic load balancing, and/or unexpected
situations like machine breakdowns and bottlenecks. Therefore, the integration of process
planning and scheduling function is established to effectively handle the dynamic circumstances.
The integration will lead to a dynamic manufacturing environment capable of reacting to
changes in the factory conditions. In fact, the integrated function will assist in producing an
optimal solution that reflects real condition on the shop floor (Wongthanasunthorn, 2007). The
integration of the two functions can introduce significant improvements to the efficiency of the
manufacturing facilities through elimination or reduction in scheduling conflicts, reduction of
flow-time and work-in-process, improvement of production resources utilization and adaptation
to irregular shop floor disturbances (Lee and Kim, 2001).
57

In this dissertation, the dynamic process planning has been successfully investigated using
computer simulation approach. To begin with, the models have been generated so as to represent
the actual manufacturing system in Arena software. There are three different models including
two static models and one dynamic model. Although the first two models are prepared to imitate
the static environment, they have been built using different methods. The first model is made
using sequence method whereas the second one is developed by variable approach. Apart from
this, the final model has achieved creating the model to deal with the dynamic circumstance in
the shop floor level. In reality, the alternative machines concept has been applied in order to
accomplish this model. The logic of this alternative route is basically that the item will have a
choice of machines to choose instead of processing through one fixed route. Arena automatically
schedules the item to the machine which presents shorter queue.
According to the first experiment, the identical outcomes between two static models have shown.
It indicates that the simulation models have been created practically. Next, the second study
proves that the manufacturing performance in dynamic process planning outperform the ones in
static planning. As alternative machines present, the manufacturing performance is improved
significantly. In alternative machine approach, not only the overall product increasing but the
time required to manufacture the item and waiting time in queue also dropping.
After that, the third experiment has arranged to observe the manufacturing performance in the
most practical shop floor environment. The result is that there is obviously one machine
presenting a very long waiting time and queue when the running time is rising. It can be said that
CNC lathe is the most preferable machine in this system and probably the one that becomes the
bottleneck. This experiment is really useful since the plan can allow the planners to know the
trouble and considers about the best possible solution. In correspondence with the last test, it
shows that increasing the number of alternative routes can apparently improve the manufacturing
performance. However, developing too many numbers of plans can possibly present the negative
outcome as you can see in the dynamic model C.
Overall, it can be concluded that an improvement of the dynamic process planning can be
achieved by using computer simulation. Alternative plans, in particular, have developed better
performance for manufacturing system only if the appropriate number of plans has been applied.
58

6.2 Recommended Further Work
Although the objectives of this research have been achieved, there are still several areas in the
topic of integration of process planning and scheduling function using computer simulation
which are recommended to work on:

To make more practical simulation experiment, other real condition data such as transfer
time, delay time, set-up time and distance between the machines should be considered.

The dynamic simulation model built in this work considers only alternative machines
approach. Developing the simulation model to cope with alternative sequences concept
could be the next step to investigate dynamic process planning in more reality.

The effects of machine breakdowns and bottlenecks in the shop floor level could be
studied further using the computer simulation. In real-time manufacturing situation, many
unexpected events can occur without caution. Thus, it is necessary to generate the plan
that can deal with these kinds of situations dynamically.

The further research should be focused on the method to generate the more precise
number of alternative plans. Due to excessive number of plans can generate negative
consequence in the manufacturing system, the best possible appropriate number of plan
should be studied.

59

List of References
Ali, Z. and Motavalli, S. (1993) Requirements from a new-generation of CAPP software. 2
nd
industrial
engineering research conference proceedings, Los Angeles, 174, pp 797-801
Alting, L. and Zhang, H. (1989) Computer aided process planning: the state-of-the-art survey.
International journal of production research, 27 (4), pp 553-585
Beckendorff, U., Kreutzfeldt, J . and Ullmann, W. (1991) Reactive workshop scheduling based on
alternative routings. In: Proceedings of a conference on factory automation and information management.
Florida: CRC Press, pp 875-885
Bhaskaran, K. (1990) Process plan selection. International journal of production research, 28 (8), pp
1527-1539
Bobrowski, P.M. and Mabert, V.A. (1988) Alternate routings strategies in batch manufacturing: An
evaluation. Decision sciences, 19 (4), pp 713-733
Bradimart, P. (1993) Routing and scheduling in a flexible job shop by tabu research. Annal of operations
research, 41, pp 157-183
Carvalho, J .D.A. (1996) An integrated approach to process planning and scheduling. Thesis, the
University of Nottingham
Chang, T.C. (1990) Expert process planning for manufacturing. Addison-Wesley Publishing Company.
Chen, C.M., Hsieh, S.T, Lin, S.T., Leslie, R., Madsen, N.H. and Beckett, R.E. (1990) CAPP for small job
shops. Proceedings of the ASME manufacturing international, New York, March 1990, pp 249-255
Chen, Q. and Khoshnevis, B. (1993) Scheduling with flexible process plans. Production planning &
control, 4 (4), pp 333-343
Chen, J . and Chen, F.F. (2008) Adaptive scheduling and tool flow control in flexible job shops.
Chryssolouris, G. and Chang, S. (1985) An integrated approach to integrate process planning and
scheduling. Annals of CIRP, 34 (1), pp 413-417
60

Conway, R.W. and Maxwell, W.L. (1962) Network dispatching by shortest operation discipline.
Operation Research, 10 (51)
Dagli, C. and Sittisathanchai, S. (1993) Generic neuro-scheduler for job shop scheduling. Computers and
industrial engineering, 25 (1-4), pp 267-270
Dong, J ., J o, H.H. and Parsaei, H.R. (1992) A feature-based dynamic process planning and scheduling.
Computers and industrial engineering, 23 (1-4), pp 141-144
Ferreira, J .C.C. and Wysk, R.A. (2000) An investigation of the influence of alternative process plans on
equipment control. J ournal of manufacturing system, 19 (6), pp 393-406
Gindy, N.N.Z., Huang, X. and Ratchev, T.M. (1993) Feature-based component model for computer-aided
process planning systems. International journal of computer integrated manufacturing, 6 (1&2), pp 20-26
Giusti, F. and Santochi, M. (1989) KAPLAN: a knowledge-based approach to process planning of
rotational parts. Annals of the CIRP, 38 (1), pp 481-484
Groover, M.P. and Zimmers, E. (1984) CAD/CAM: Computer-aided design and manufacturing.
Englewood Cliffs, NJ :Prentice-Hall
Gu, P. and Zhang, Y. (1994) OOPS: An object-oriented process planning system. Computers and
industrial engineering, 26 (4), pp 709-731
Halevi, G. (1980) Development of flexible optimum process planning procedures. Annals of the CIRP, 29
(1), pp 313-317
Hormazi, A.M. (1994) Agile manufacturing. American Production and Inventory Control Society, pp
216-218
Huang, S.H., Zhang, H.C. and Smith, M.L. (1995) A progressive approach for the integration of process
planning and scheduling. IIE Transactions, 27, pp 456-464
Iwata, K. and Fukuda, Y. (1989) A new proposal of dynamic process planning in machine shop. The
CIRP International Workshop, Hannover, Germany, September, pp 104-109
Iwata, K., Fukuda, Y. and Sugimura, N. (1991) A proposal of knowledge structure for integrated process
planning system. Manufacturing system, 20 (3), pp 273-282
61

J oo, J .K., Park, S.S. and Cho, H.B. (2001) Adaptive and dynamic process planning using neural networks.
Int. J . Prod. Res., 39 (13), pp 2923-2946
J oshi, S., Vissa, N.N. and Chang, T.C. (1988) Expert process planning system with solid model interface.
Kelton, W.D., Sadowski, R.P. and Sturrock, D.T. (2003) Simulation with Arena. 3
rd
edition, Singapore;
McGraw-Hill
Kempenaers, J ., Pinte, J ., Detand, J . and Kruth, J .P. (1996) A collaborative process planning and
scheduling system. Advances in Engineering Software, 25 (1), pp 3-8
Khoshnevis, B. and Chen, Q.M. (1990) Integration of process planning and scheduling function. J ournal
of intelligent manufacturing, 1, pp 165-176
Khoshnevis, B. (1990) Integrated process planning. Proc. Of manuf. International 90, ASME, Atlanta,
Georgia, March 25-28, pp 243-248
Kim, K., Park, K. and Ko, J . (2003) A symbiotic evolutionary algorithm for the integration of process
planning and job shop scheduling. Computer Operation Research, 30, pp 1151-1171
Kuhnle, H., Braun H.J . and Buhring, J . (1994) Integration of CAPP and PPC-interfusion manufacturing
management. Integr. Man. Syst., 5 (2), pp 21-27
Kumar, M. and Rajotia, S. (2003) Integration of scheduling with computer aided process planning.
J ournal of Materials Processing Technology, 138, pp 297-300
Larsen, N.E. and Alting, L. (1992) Dynamic planning enriches concurrent process and production
planning. Int. J . Prod. Res., 30 (8), pp 1861-1876
Larsen, N.E. (1993) Methods for integration of process planning and production planning. International
J ournal of Computer Integrated Manufacturing, 6 (1-2), pp 152-162
Lee, H. and Kim, S.S. (2001) Integration of process planning & scheduling using simulation based
genetic algorithms. International J ournal of Advanced Manufacturing Technology, 18, pp 586-590
Lim, M.K. and Zhang, Z. (2003) A multi-agent based manufacturing control strategy for responsive
manufacturing. J ournal of Materials Processing Technology, 139, pp 379-384
62

Lim, M.K. and Zhang, D.Z. (2004) An integrated agent-based approach for responsive control of
manufacturing resources. Computers & Industrial Engineering, 46, pp 221-232
Ming, X.G., Mak, K.L., Yan, J .Q., Ma, D.Z. and Zhang, H.Q. (1997) A hybrid expert neural-based
function model for CAPP. International J ournal of Computer Integrated Manufacturing, 10, pp 105-116
Nahmias, S. (2005) Production and operations analysis. 5
th
international edition, Boston; London:
McGraw-Hill Irwin
Nwana, H. and Ndumu, D. (1997) An introduction to agent technology. Software agents and soft
computing: toward enhancing machine intelligence, concepts and applications, pp 3-36
Pidd, M. (1992) Computer simulation in management science. 3
rd
edition, West Sussex, England; J ohn
Wiley & Sons
Pinedo, M. and Chao, X. (1999) Operations scheduling with applications in manufacturing and services.
International edition, Irwin/McGraw-Hill
Rembold, U., Nnaji, B.O. and Stor, A. (1993) Computer integrated manufacturing and engineering,
Addison-Wesley, ISBN 0-201-56541-2
Robinson, S. (1994) Successful simulation A practical approach to simulation projects. McGraw-Hill
Book Company, London, England
Rodammer, F.A. and White, K.P. (1988) A recent survey of production scheduling. IEEE transactions on
systems, man, and cybernetics, 18 (6), Nov/Dec
Saygin, C. and Kilic, S.E. (1999) Integration flexible process plans with scheduling in flexible
manufacturing systems. Integration J ournal of Advanced Manufacturing Technology, 15, pp 268-280
Schmidt, B.C. and Kreutzfeldt, J . (1992) Concurrent process planning and scheduling using non-linear
process plans. Flexible automation and information management
Shao, X., Li, X., Gao, L. and Zhang, C. (2008) Integration of process planning and scheduling a
modified genetic algorithm based approach. Computers & Operations Research, doi: 10.1016/
j.cor.2008.07.006
Shen, W.S., Hao, Q., Yoon, H.F. and Norrie, D.H. (2006) Applications of agent-based systems in
intelligent manufacturing: An updated review. Advanced engineering informatics, 20, pp 415-431
63

Song, L., Wang, P. and AbouRizk, S. (2006) A virtual shop modeling system for industrial fabrication
shops. Simulation modeling practice and theory, 14, pp 649-662
Srihari, K. and Greene, T.J . (1988) Alternative routings in CAPP implementation in FMS. Computers &
industrial engineering, 15 (1-4), pp 41-50
Sundaram, R.M. and Fu, S.S. (1988) Process planning and scheduling a method of integration for
productivity improvement. Computers & industrial engineering, 15 (1-4), pp 296-301
Tan, W. and Khoshnevis, B. (1997) Integrated process planning and scheduling a mathematical
programming modeling approach, working paper 1997-02, Department of Industrial & Systems
Engineering, University of Southern California
Tan, W. and Khoshnevis, B. (2000) Integration of process planning and scheduling a review. J ournal of
Intelligence Manufacturing, 11, pp 51-63
Tonshoff, H., Beckendorff, U. and Andres, N. (1989) FLEXPLAN: A concept of intelligent process
planning and scheduling. The CIRP International Workshop, Hannover, Germany, September, pp 319-
322
Tonshoff, H., Kreutzfeldt, J . and Hofschneider, D. (1992) Concurrent process planning and workshop
control in batch production, Proc 24
th
CIRP intern. sem. on manu. systems, 11-12, Copenhagen, Denmark,
pp 315-328
Usher, J .M. and Fernandes, K.J . (1996) Dynamic process planning the station phase. J ournal of
Materials Processing Technology, 61, pp 53-58
Usher, J .M. (2003) Evaluating the impact of alternative plans on manufacturing performance. Computers
& Industrial Engineering, 45, pp 585-596
Wong, T.N. and Siu, S.L. (1992) An automated process planning system for prismatic parts. Int. J . of
computer integrated manufacturing, 5 (4-5), pp 319-333
Wong, T.N., Leung, C.W., Mak, K.L. and Fung, R.Y.K. (2006) Integrated process planning and
scheduling/rescheduling an agent-based approach. International J ournal of Production Research, 44 (18-
19), pp 593-605
64

Zhang, H.C. (1993) IPPM-a prototype to integrated process planning and job shop scheduling functions.
Annals of the CIRP, 42 (1), pp 513-518
Zhang, H.C., Huang, S.H. and Mei, J . (1993) An introduction to an integrated and intelligent process
planning model. ASME Manufacturing Science and Engineering, 64, pp 291-298
Zhang, Y.F., Saravanan, A.N. and Fuh, J .Y.H. (2003) Integration of process planning and scheduling by
exploring the flexibility of process planning. Int. J . Prod. Res., 41 (3), pp 611-628
Zhang, W.J ., Xie, S.Q. (2007) Agent technology for collaborative process planning: a review.
International J ournal of Advanced Manufacturing Technology, 32, pp 315-325

65

Appendices

66

Appendix A: Comparison of Integration Models

Advantages Disadvantages
NLPP
Providing all the alternative process plans
of, and enhancing the flexibility and the
availability of process plans.
Because of the need to give all alternative
process plans of the parts, this will cause a
combinational-explosive problem.
CLPP
Based on the current shop floor status, the
process plans are all very useful.
CLPP needs the real-time data of the current
status, if it has to re-generate process plans in
every scheduling phase, the real-time data is
hard to be assured and updated.
DPP
This model works in an interactive,
collaborative, and cooperative way
Because the basic integration principle of DPP
is a hierarchical approach, it cannot optimize
the process plans and scheduling plans as a
whole.

Table A1: Comparison of integration models <Shao et al, 2008>

67

Appendix B: Static Model Using Sequence Method

Figure B1: Overall model of the static model using sequence method.

68

Figure B2: A detailed of Create module for item 1

69

Appendix C: Static Model Using Variable Method

Figure C1: Complete model of the static model using variable method.

70

Figure C2: The Variable data module of the static model using variable method. It includes the
detailed of Optimes variable in the array window.

71

Appendix D: Dynamic Model

Figure D1: Completed model of the dynamic model using variable method

72

Figure D2: The Variable data module of the dynamic simulation model. It consists of the
detailed of initial values applied to Optimes variable.

Figure D3: The Variable data module of the dynamic simulation model. It consists of the
detailed of initial values applied to QLS variable.
73

Figure D4: A detailed of Decide module from the Submodel. The condition is to decide whether
the item has a choice of machine or not.

Figure D5: A detailed of second Decide module in Submodel. Item will choose the machine
with shorter queue to be its next destination according to the defined Variable data module.

74

Appendix E: Table of Alternative Routes for Experiment 2 & 3

OP 1 (M/C, minute) OP2 OP3 OP4 OP5 OP6 OP7 OP8
Item 1 1,1 2,5 5.1,4 5.2,3 6.1,6 3.1,12
(6,1) (1,10)
Item 2 1,2 2,15 5.1,5 5.2,2 6.1,3 7,2
(5,7) (2,10) (2,5)
Item 3 1,3 2,30 5.1,20 5.2,15 6.1,20 6.2,15 6.3,20 7,10
(3,30) (3,20) (3,25)
Item 4 1,3.5 2,10 5.1,3 5.2,4 6.1,8 6.2,7
(5,4) (2,7) (2,7)
Item 5 4.1,8 4.2,8 4.3,5 6.1,15 6.2,10 7,8
(6,3) (6,4) (6,2) (3,40) (3,20)

Table E1: Table representing the processing times for different items includes the introduction
of alternative machines. This is used as data for generating the dynamic model.
Where:
Item 1: Flexible pin base Machine 1: Cutting
Item 2: Bearing mount Machine 2: Lathe (2P)
Item 3: Motor stand Machine 3: Milling (2P)
Item 4: Slide bush Machine 4: Milling face (3P)
Item 5: Wheel holder Machine 5: CNC lathe (2P)
Machine 6: CNC milling (3P)
Machine 7: Tapping

75

Appendix F: Table of Alternative Routes for Experiment 4
OP 1 (M/C, time) OP2 OP3 OP4 OP5 OP6 OP7 OP8
Item 1 1,1 2,5 5.1,4 5.2,3 6.1,6 3.1,12
(6,1) (4,5) (1,10)
Item 2 1,2 2,15 5.1,5 5.2,2 6.1,3 7,2
(5,7) (2,10) (2,5) (1,6)
Item 3 1,3 2,30 5.1,20 5.2,15 6.1,20 6.2,15 6.3,20 7,10
(4,20) (3,30) (3,20) (3,25)
Item 4 1,3.5 2,10 5.1,3 5.2,4 6.1,8 6.2,7
(5,4) (2,7) (2,7) (1,12)
Item 5 4.1,8 4.2,8 4.3,5 6.1,15 6.2,10 7,8
(6,3) (6,4) (6,2) (3,40) (3,20) (1,8)

Figure F1: Table representing the sequence of items along with processing time. This is used to
build the dynamic model A.

Item 1 1,1 2,5 5.1,4 5.2,3 6.1,6 3.1,12
(6,1) (4,6) (4,5) (1,10)
Item 2 1,2 2,15 5.1,5 5.2,2 6.1,3 7,2
(5,7) (2,10) (2,5) (1,6) (1,2)
Item 3 1,3 2,30 5.1,20 5.2,15 6.1,20 6.2,15 6.3,20 7,10
(4,25) (4,20) (3,30) (3,20) (3,25)
Item 4 1,3.5 2,10 5.1,3 5.2,4 6.1,8 6.2,7
(5,4) (2,7) (2,7) (1,12) (4,10)
Item 5 4.1,8 4.2,8 4.3,5 6.1,15 6.2,10 7,8
(6,3) (6,4) (6,2) (3,40) (3,20) (1,8)

build the dynamic model B.

76

77

Item 1 1,1 2,5 5.1,4 5.2,3 6.1,6 3.1,12
(6,1) (4,5) (4,6) (4,5) (1,10) (6,5)
Item 2 1,2 2,15 5.1,5 5.2,2 6.1,3 7,2
(6,1) (5,7) (2,10) (2,5) (1,6) (1,2)
Item 3 1,3 2,30 5.1,20 5.2,15 6.1,20 6.2,15 6.3,20 7,10
(7,2) (3,15) (4,25) (4,20) (3,30) (3,20) (3,25) (1,2)
Item 4 1,3.5 2,10 5.1,3 5.2,4 6.1,8 6.2,7
(7,2) (5,4) (2,7) (2,7) (1,12) (4,10)
Item 5 4.1,8 4.2,8 4.3,5 6.1,15 6.2,10 7,8
(6,3) (6,4) (6,2) (3,40) (3,20) (1,8)

build the dynamic model C.

08MScLIXPU1 2

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

08MScLIXPU1 2

Uploaded by

Copyright:

Available Formats

INVESTIGATION OF DYNAMIC

PROCESS PLANNING USING

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul

Investigation of Dynamic Process Planning using Computer Simulation Arena P.Umnutkittikul