Professional Documents
Culture Documents
The Greater China Electronics Industry - An External View Pages 24-29
The Greater China Electronics Industry - An External View Pages 24-29
The Greater China Electronics Industry - An External View Pages 24-29
1996-97
CONTENTS
Message from Dr. Alan H.S. Chan,
President, HE (HK) ..................................................................................................1
Message from Mr. John J. Powers
President, HE ............................................................................................................2
Message from Prof H.K. Chang
President, City University of Hong Kong .................................................................3
Message from Prof Qi Ershi
Executive Director, Industrial Engineering Institution of China ...............................4
Publisher
Directors, Advisors &Corporate Members 1996-97 ..............................................6
Institute of Industrial Engineers (HK) Ltd.
Office Bearers and Executive Committee 1996-97 ..................................................7
Activities and Collaborated Events for 1996-97 ......................................................8
Tenth Scholarship Presentation ................................................................................9
Institute of Industrial Engineers (Hong Kong) .........................................................10
Editors
Membership Application Form ...............................................................................12
Dr. L.K. Chu
Papers
Dr. Alan H.S. Chan
Advanced Design & Manufacturing Technologies
A Virtual Reality-based Training System for CNC Milling Machine Operations
F. H. Lin, C.L. Hon & CJ Su ...................................................................................13
IIE (HK) Ltd. accepts no responsibility to connection
Experimental Study of a Solder Paste Stenciling Process using Factorial
with any liability which might develop as a result of Design
material published; opinions expressed are those of G.K K Poon ..............................................................................................................17
the authors and do not necessarily represent the Application of Structured Analysis in the Product Creation Process
Institute. C K. Kwong & W. C. Wong ....................................................................................20
Global Manufacturing
The Greater China Electronics Industry - an External View
K J Tilley& D. J. Williams .......................................................................................24
The Institute of Industrial Engineers (HK) Manufacturing Systems
Limited by Guarantee Application of Genetic Algorithms in Optimizing Parameters in Decision
G.P.O. Box 6635, Hong Kong Support System for Resource Allocation
K L Choy & V. W. S. Yeung.......................................................................................31
The Design of a Generic Flexible Manufacturing Cell
CKF. Chiu& L.K Chu............................................................................................... 35
Service Industries
Front cover design Service Productivity & Quality Excellence in the Air Express Industry
Mr. Sam Deng YM Wong ..................................................................................................................39
Industrial Engineering Education
Industrial Engineering in Hong Kong
R.C Whirfeld ..............................................................................................................45
Printing Member List .............................................................................................................48
Media Production Unit Companies Employing IIE (HK) Members ............................................................55
City University of Hong Kong
To the IIE, this year is one of consolidation and growth. In recent times we have seen a
steadily increasing membership, which has grown 20% in the past year, and I believe this
growth will continue in the future. It is hoped that the membership directory at the back of the
Annual Journal provides a vital service to the members.
With the changing nature of industries, being an Industrial Engineer is more challenging today than in the past. These
changes are strongly stimulating the need for competent professionals with IE knowledge and skills, both in Hong Kong and in
China. The HE (HK) recognizes these chan ges and challenges and has decided to institute close partnerships with IEs in China
by increasing the level of interaction and improving the communications between professionals in these two regions. An
Agreement has been reached between us that permits more intense communications through the interflow of publications, and
seminars and conferences, as will be seen in the coming years. In this, we are taking the first step forward by jointly organising
the 5th Industrial Engineering Conference with the Industrial Engineering Institution in China next year.
Certainly, we need to satisfy the needs of local industries which have large manufacturing and operational bases across
the border. In co-operation with the Federation of Hong Kong Industries, the Institute is now beginning to organise Industrial
Engineering related training activities for engineers in Shenzhen. Accreditation of these courses by the HE Headquarters is
being sought, so they can be used for membership criteria.
1997 is not only meaningful for all Chinese, but is particularly memorable for the IE profession in Hong Kong, as the
third IE undergraduate program is to be launched in the City University of Hong Kong next year. This is in addition to the two
existing courses in the Hong Kong University, and the Hong Kong University of Science and Technology. We would like to
see closer collaboration between the Institute and these three Universities to foster the discipline of Industrial Engineering in
the coming years.
One of the objectives of the Institute is to promote the dissemination of knowledge and information by publications
relating to the art and science of Industrial Engineering. This year we are taking the initiative to have our own
refereed international journal "Industrial Engineering Research" which will report basic and applied research in
Industrial Engineering. I am pleased to report that after months of preparation work, we have built a strong editorial
team with distinguished scholars from the UK, China, Australia, Taiwan, the USA and Hong Kong.
Seminars, workshops, membership drives, and technical visits have also been kept going to provide continuous services
and professional development opportunities for our members to meet their professional needs.
I would like to take this opportunity to thank the Executive Committee this year, who have made tireless efforts to ensure
the success of the Institute. Last but not least, I would like to thank the Institute's Directors, Corporate Members, and
scholarship sponsors for their continuous advice and support, without which our continuous and prosperous growth would not
have occurred.
As we move toward the 21 st century, all industrial engineers in Hong Kong face
increasing opportunities. The continued development of the Pacific Rim and the ongoing
multi-national economy bring about an unprecedented series of events that demand the
expertise, analytical tools, design methodologies, and management skills that industrial
engineers uniquely possess.
The Institute of Industrial Engineers is proud to give 100% support to the Hong Kong
Chapter of he Institute of Industrial Engineers, and the worldwide profession of industrial
engineering. As industrial engineers in Hong Kong step up to the opportunities before them,
all 26,000 members of the Institute stand with them for continued success.
John J. Powers
President, 1996-97
Institute of Industrial Engineers
The Institute of Industrial Engineers (HK), with the active participation of its members,
has helped Hong Kong's industries to overcome these challenges in economic transition.
Together, they have made full use of engineering and re-engineering processes and systems to
achieve better quality and higher productivity during product development. The latest changes
also offer good opportunities for application of advanced industrial engineering
methodologies and techniques.
On this 18th Anniversary of the Institute of Industrial Engineers (HK), I would like to
congratulate the Institute for its remarkable achievement in the past and I wish the Institute
and its members every success in the future.
Executive Committee
Engineering Mr. Louis Tam Mr. S.H. Lee
GE - LG Appliances Avery-Dennison
Internet Services Mr. Ricky Yeung Mr. H.F. Wong
City University of Hong Kong Team Concept Ltd.
Membership & Mr. Eric Choi Mr. M.W. Lam
Scholarship Hong Kong Telecom Avery-Dennison
Publications Dr. L.K. Chu Dr. H.W. Law
The University of Hong Kong City University of Hong Kong
Social Activities Mr. Hudson Cheung Mr. Wallace Lam
Hong Kong Government Hong Kong Technical College (Tsing Yi)
Officers Ms Iris M.Y. Chan Mr. James Leung
Mr. Stefan Chow Ms Kelly S.L. Liu
Mr. Ricky Fung Mr. Devip Yip
Mr. H.M. Leung Mr. Andy Yeung
Newsletter: Number 1
Newsletter: Number 2
NOV 2nd Internet Workshop - Internet Phone, Fax, chat and similar Internet services
Symposium on Global Manufacturing
Newsletter: Number 3
DEC Seminar on Manufacturing Technologies
1997
JAN 3rd Internet Workshop - Introduction to HTML and homepage language
Industrial visit
Newsletter: Number 4
Industrial visit
Newsletter: Number 6
Since the start of the Scholarship Program, we have unnoticeably been moving into the first decade. In this year, fourteen
outstanding full-time students of Industrial Engineering (or equivalent) curriculum from six local educational institutes are
awarded the 10th IIE (HK) Scholarship after a stringent selection process. On behalf of the HE (HK), may I take this
opportunity to congratulate the awardees and wish them every success in their academic and career path.
Although a lot of companies in Hong Kong are now suffering from the global economic recession and $ome of them have
started the first step of downsizing, our career as an Industrial Engineer is glowing. In fact, it provides us a playground to
demonstrate our intrinsic cap ability of designing and carrying out productivity and quality improvement projects established
during our studies.
Finally, I would like to cordially thank our generous sponsors of the Scholarship Program and all the faculties/
departments ofthe six institutes who furnish us with their nominations and assistance and I am looking forward to a more closer
collaboration with all of them in the future.
Membership Qualifications
• Graduate of Industrial Engineering (or equivalent) curriculum or option, accredited by the accrediting agency
recognized by the HE : 5 years IE experience for senior member (Note a).
• Graduate of other engineering curriculum accredited by the accredit ing agency recognized by the IIE, graduate
of a masters or doctoral degree program administered by a department offering an IE curriculum accredited by
the accrediting agency recognized by the HE or graduate of an IE curriculum not accredited at time of
graduation but which becomes accredited or reaccredited within two years following date of graduation : 2 years
IE experience for member and 7 years for senior member (Note a).
• Graduate of engineering curriculum (other than IE) in the jurisdiction of the accrediting agencies recognized by
the HE or other baccalaureate degrees, including technology degrees other than accredited IE Technology: 5
years IE experience for member (Note b).
• Any applicant having less than a 4-year baccalaureate degree (Note c) : 8 years of IE experience for member
(Note b).
Note a
Two years of credit will be allowed for being a Registered Professional Engineer.
Note b
Eligible if applicant qualifies for Member grade and is either a Registered Professional Engineer or with adequate justification in writing from the chapter (or
individual if member-at-large) upon recommendation by the Regional Vice President and approved by the Institute Board of Review.
Note c
Appropriate credit will be allowed, up to a maximum of two years, for college level engineering or engineering courses satisfactorily completed.
system prompt "Manual operation "when mouse flip it `B' Set the depth (Z-axis) of 1 st hole will be drilled
`C' Separate the mill and workpiece after 1 st hole have been
system will clear the previous enter or escape the next drilled
`D' Set the position of 2nd hole will be drilled at (X=1.75,
Y=-0.75, Z=0.0)
These number keys are the numerical key and
the user can key in the mini or decimal number `E' Set the depth (Z-axis) of 2nd hole will be drilled
when the key is flipped.
`F' Drill the 2nd hole and finish all the training procedures
Under this process the user can remember the routine operation
procedures easily and avoid making the mistakes to a lower extent.
Roughly, there are four operating stages: initializing the machine,
setting- up working zero point, milling, and shutting- down the
machine. After training for setting up, the user will know the basic
skills of CNC milling operations: (1) how to initialize the CNC
machine; (2) how to set the mill to working zero posit ion. For
milling operations, the main issue to be studied is to determine the
possible scenarios which may occur during the real operation. After
this, we can design the guiding strategies and evaluating plans for
two training modes in our system including 1) Guide-learning Mode
and 2) Self learning Mode. In the first mode, the user is restricted to
follow the system instruction to perform task given. Figure 2
illustrates the procedure of this mode. In the second mode, the user
operates the machine by using the knowledge and experience learnt
in the Guide-learning mode. The system would not prompt what he
or she should do next during the training period, but would record
the mistakes of operation made by the user. After finishing all
training procedures, the CRTS system will generate an evaluation
report that contains the information such as the number of mistake
Fig 1 The three-axis CNC milling machine model, workpiece and a made, the operations which the users need to practice further, and so
control panel on. Figure 3 illustrates the procedure of this mode.
3 Scenarios Analysis
Our goal is to teach the trainee how to operate the machine in 4 Collision Detection
manual mode so that after the virtual training the user can machine a As a foundation of interaction, collision detection between the
simple part in the real machine. First we surveyed all the possible milling machine and other objects such as electronic probe and
scenarios in the fifteen basic tasks in milling machine operation and virtual hand is necessary in the training system. To enable and
the related buttons necessary in manual operation. Then we defined a trigger the response, such as the motion constraints of objects, when
training process by ignoring unnecessary buttons on the control panel objects in the VRTS are colliding with each other,
and procedures not
5 Experiments
We used SenseB's WorldToolKit 2.1 for PC and UNIX as main
developing tools. All the objects were built by using 3D Studio 4.0 in
PC. These objects models were then converted to WorldToolKit for
SGI Onyx format. The system runs on 100MHz PCs with Pentium
processors and 16Mb RAM or SGI ONYX. Polhumus Fastrak,
CrystalEyes glasses and Emitter were used to get stereoscopic effect.
The device controllers for Cyberglove and mouse were also
developed for the user- system interaction in our VRTS. The key
components of the system include: a 3D geometrical model of the
CNC machine, a guide training module, a self-learning module, an
infra-communication control modules, an interactive menu modules,
a collision detector, and a performance evaluator.
6 Conclusion
In this paper, a virtual reality-based training system (VRT S) for
CNC milling machine operations is described. The developed virtual
environment has the following advantages over the traditional train
methods: (1) the system provides safe and economical training
environment; (2) the training system provides two different training
patterns for different purposes; (3) on- line interactive instructions
ensures that the correct tasks are effectively carried out by "virtually"
operating; (4) the system provides the flexibility to structure training
programs and even monitors and measures the progress of a training
session. (5) it is possible that the users can be trained in the remote
site through Internet by using this computer software.
References
1. K. Warwick, J. Gray & D. Roberts, VIRTUAL REALITY
ENGINEERING, IEE computing series, 1993.
5 Conclusion
The effects of several parameters of a solder paste stenciling
process on the printing quality is studied effectively using factorial
experim ents. The temperature of the solder paste and cleaning
interval of the metal squeegee are found to have significant negative
effect on the PVM and NPU of the printed boards, and hence a low
level of both factors is recommended. Moreover, the noise effect of
the paste viscosity on the PVM is also found to be minimized when a
low squeegee pressure and cleaning interval is used. A linear
prediction model is established and verified to be highly accurate
within the experimental region.
References
1. J.S. Hwang, SOLDERPASTE IN ELECTRONIC PACKAGING,
NY Van Nostrand Reinhold 89.
2. S.K. Buttars, "Parameters of Solder Paste Printing for Fine Pitch
Components", Proc. NEPCON-WEST `93, pp 12541265.
3. J.S. Hwang, "Solder/Screen Printing", Surface Mount
Technology, March 94, pp44-50.
4. N.N. Ekere, E.K. Lo & S.H. Mannan, "Process Modeling Maps
for Solder Paste Printing", Soldering & Surface Mount
Technology, No. 17, May 94, pp4-11.
5. J.R. Morris & T. Wojcik, "Stencil Printing of Solder Paste for
Fine-pitch Surface Mount Assembly", Soldering & Surface
Mount Technology. No. 5, 1990, pp10-14.
6. I. Ismail, S.H. Mannan, N.N. Ekere & E.K. Lo, "Experimental
Study of the Printing of Solder Paste using the Metal Blade
Squeegee System", Proc. of the 30th Int. MATADOR Conference,
UMIST 1990, pp263-266.
7. NN. Ekere, E.K. Lo & I. Ismail, "Experimental Study of Stencil
Substrate Separation Speed in Solder Paste Printing for ReAow
Soldering", Journal of Electronics Manufacturing, Vol. 3, No. 1,
April 93, pp25-29.
processes and outputs are re-constructed in a systematic format. The Apart from identifying the bottlenecks, the IDEFQ models can
Product Creation Process is a structured and systematic project also help to identify the problems in employing different design tools
management system. It governs the orders of functions as well as for dissimilar design applications within the Product Creation Process.
controlling the input and output specifications. The system could be With reference to Table 1, it is found that different functions are
further decomposed into a functional hierarchy of activities as shown adopting different design tools for their jobs.
in Figure 3. Each functional hierarchy can be further broken into The use of different design tools between similar applications
more detailed activity models by using IDEFo. Figure 4 shows the does not cause too much troubles. Simple transformation on data
IDEFo activity model of Function A4 (Product Design Details). type and/or media types may help to bridge the gap. For example,
switching from AutoCAD to ClarisCAD for 2D CAD applications
With the use of the IDEFo models, five different functions in are relatively simple, since the data structures of both applications
are similar, even though it requests addition efforts and it may
reduce the reliability of the data obtained.
Work transfer between dissimilar applications always create
problems. For example, porting data between 2D CAD and 3D CAD
applications, or even transferring data between different 3D CAD
applications may lead to unpredictable data errors.
6 Analyzing The Information Flow By Using Data
Entity Matrix
The Product Creation Process involves the use of manual
systems as well as computer-aided systems. Information
transformations are not only limited to electronic data transfer, they
also include paper documents and hand sketches. As a high degree
of human judgment is involved throughout the Product Creation
Process, the flow of information is complicated. If it is not well
structured, the efficiency of the entire system will be reduced.
In order to identify areas of ambiguous, information
redundancies and any non-value added functions in the information Make Working Prototype (A54) Manual/ MasterCAM
transformation process, a simple Data Entity Matrix is employed. The Construct 2D & 3D Geometry Data (A621) Manual / CADDS / Sicignaph,
m atrix is constructed as an extension of the IDEFo activity model. It
UGH
focuses on the transformation of information between 2D and 3D
data, and the effects caused by Table 1 Design Tools Employed in Various Functions
7 Conclusion ,
Shortening the product design and
prototype development cycle is essential
for improving the `time-to-market' value.
With the use of computers, it is hoped
that the quality of work of the product
designers can be improved and so the
lead-time for product development can
be shortened. Unfortunately, as human
judgment is heavily involved in the
product development process, frequent
human errors in data interpretation
highly skill dependent. It starts with function A421, which is a and mismatching in data type conversion occur. To harmonize the
reading and an interpreting process based on 2D information on data and information flow, structural methods (IDEFo and Data
model drawings. Then, the model maker bases on his perceived 3D Entity Matrix) are adopted to analyse the existing information system
image to construct the model. The designer's concept and the model throughout the design process.
maker's perceived 3D image may deviate from each other due to the With the use of the IDEFo model, the bottlenecks in the Product
following possible reasons: Creation Process can be identified. Resources can then be
re-deployed to eliminate the bottlenecks and hence the product
• the quality of information on the model drawings,
development cycle time reduced.
• the experience level on constructing 3D objects, The Data Entity Matrix is then adopted as an extension of the
• the background and the interest of the designer, as well as IDEFo activity model to identify the degree of human involvement,
that of the model maker, and areas of data ambiguous and any non-value added activities within
the Product Creation Process. Based on the findings from the IDEFO
• human intuition and knowledge. Activity Model and the Data Entity Matrix, problems in the existing
Since the interpretation process is not a formal procedure and system can be identified and directions for improvement can be
may not be well documented, problems and errors may not be proposed.
identified easily in the first place. Such hidden errors may cause
confusion in successive functions. References
Actually, the process of data transformation can be described as 1. D.A. Marca, IDEF / SADT BUSINESS PROCESS AND
a mechanism to change the state of data, from its input state to its ENTERPRISE MODELLING, Eclectic Solutions Corp., 1993.
output state. The process can be a straightforward machine function, 2. B.T. Brooksby, "Design Process Improvement", IEE
Colloquium, n 223, 1994, p.2/1-5.
or it can be a complex human interpretation that heavily influences
the efficiency of exchange, as well as the quality of output. Once the 3. J.R. Moyne, L.C. McAfee & T.J. Teorey, "Application of Entity
area and context of human involvement are identified, the reliability Relationship Data Modelling Techniques to the Automated
and significance of data exchange could be verified. Hence Manufacturing Process", Proc. of 2nd Int'1 Conf on Data and
streamlining the data exchange between 2D state and 3D state knowledge Systems for Manufacturing and Engineering, 1989,
becomes possible. pp206-215.
In order to improve the quality of the design data and shorten the 4. R.G. Ross, ENTITY MODELLING: TECHNIQUES AND
product development cycle, the manual- driven design system has to APPLICATION, Database Research Group, 1988.
be modified to fit with the different automated design tools used.
Common input and output data formats have to be adopted. This
ensures design information generated by the preceding process is in
the same format as the information used by the succeeding process.
Design data can then be used
Acknowledgements
Thic x~ vas teed bythe EPSRC under grant GR/K 49768
Abstract genes. Stoppage of iteration would occur if the difference of the top
Genetic algorithms (GA), since first introduced by Holland [1] in results in the last successive K operation are within 5% of the total
early 1970s, was applied on optimization problems and showed variation range. K can be changed to suit the number of operation
significant performance. However, it was seldom mentioned in cycle. By using the above technique, optimal parameters for a
manufacturing literature, especially in optimizing resources production system can be determined which could be used for
allocation problems and line balancin g problems. This paper is problems such as in line balancing and resources allocation.
concerned with applying the GA technique, interfacing with the
simulation software, WITNESS, to perform experiments in order to 2 Basics of Genetic Algorithms
obtain a set of parameters for determining the resource level in GA's involve an imitation of nature, which include goals to
setting up a production plant. By using this technique, optimal survive, population to live, mating, reproduction and mutation [3].
parameters for a production system can be determined which will Usually, the crowd of specifications will be represented by a definite
increase the effectiveness of resources utilization. The method is to amount of configurations, where each of them acts as one living thing
corporate GA's for optimizing parameters using C++ programming in nature, called Structure (chromosomes or species in nature). They
techniques, and then applying simulation as the optimizing join together to imitate a world called population. However, the
parameters evaluation tools. By using this techniques, optimal population usually contains a large number of structures or even in
parameters for a production system can be determined and beneficial expanding size in order to maintain randomization.
to resource allocation and eventually order taking of the organization. These structures consist of fundamental units called bit, position,
or token (genes in nature) which makes them different from each
1 Introduction other. A unit is one coding showing the state of a parameter.
Genetic algorithms involve converting experiment parameters Therefore, one parameter of the problem may be represented by one
into genes, which can be systematically recognized by computer. or more units.
For some simple parameters, say only yes or no, can be simply Crossover
represented by genes of 0 and 1. Some more complicated In the crossover process, two species are selected. Points on
parameters can be alphabets other than binary, or may have a rather chromosomes are then selected to break and exchange. After that two
different method in conversion. These genes of one combination of new offspring will be produced. Theses process repeats until all
parameters form a set called chromosome or string. Other different species, required to be cross overed, are cross overed. The proportion
parameters will form other different chromosomes, and all of the cross overing species to population size is the crossover rate.
chromosomes represent a population. The quality of the results will Hence, in one crossover iteration, { [crossover rate] x [population
be the guideline of determining capability of the chromosomes, and size] } structures are cross overed.
is so called fitness. The point to terminate GA's iteration is different Mutation
for different characteristic of usage. However, with small crossover Mutation is another development of chromosomes. Unlike
rate and mutation probability, the trends of most optimum output of crossover, it does not rely on mating, but happen on individual gene.
fittest chromosome in each GA operation are similar [2]. Mutation is a sudden change of gene, leading to change of
The method used here is to corporate GA's for optimizing chromosome. Occurrence of mutation is random. In GA's, we can
parameters using C++ programming technique, and them applying select randomly which gene to change to a random value.
simulation using WITNESS as the result parameters evaluation Mutation rate is the probability that an individual fundamental
tools. A random number generator for experimental parameters unit would perform mutation. Hence, in one iteration, {[m utation rate]
would also be constructed and utilized for construction of initial x [structure length] x [population size]} tokens are mutated.
population. When the first program is finished, it would be required From the above philosophy, it can be seen that GA's would
to validate the method of GA's on manufacturing optimization. Here, provide better and better results. After all, there will be saturation, or
the parameters are the inputs to simulation model, and each optimum approach.
changeable parameter behaves as a piece of gene.
The optimization process using GA operators would choose 3 Application of Genetic Algorithms to
80% of the population, in the order of the best first, and randomly Manufacturing
select pair to crossover. The remaining would be mutated in the GA's, since first introduced by Holland in early 1970s, was
probability of 0.5%, changing the value of the genes randomly applied on optimization problems and showed significant
within the constraint. The new genes would be input to WITNESS performance. Nowadays, there is annual official conferences to
for analyzing the performance. The process will continue with the study this algorithms mainly in the application of GA's and
new set of genes being created using the genetic algorithms, fed to
the simulator for performance and then send the result to analyze
using GA's to create other set of
Overall System
In the whole system, a goal must be defined. To simplify
the experiments and programming, a single goal is temporally
used and is retrieved directly from the simulation report files
(CSV files). However, the aim of optimization may be to
maximize, minimize or approach a preset value.
Another set of population of space, the controls is reserved
in this system. This space can be varied in the source code of
program. This space are slots for structures which are the
fittest in recent iterations. If the best fit structure obtains better
fitness than the worst fit one in the control population during
each experiment, the former will take the later place. While
the iterations show declining trend in fitness, all the control
structures will be copied to replace the weakest in the
processing population. The purpose of the process is to retain
the best structures comparative to the history, and to ensure
the inclining results. The reason for not copying all the better
structures to the control population is to avoid rapid
convergence of results, which may lead to termination of
program before optimum may occur.
Termination of iterations depends on the range of controls
and total range of results in history. Two variables, mainly
result min and result _max are defined such that they store the
Fig 3 Logic of interfacing G.A. Shell with WITNESS minimum and maximum value that the structures ever had
since
limit the amount from two to five the first iteration. During the iteration, if new maximum or
machines. new
The variables may also be local variables and lg obal minimum occurs, the range between result-min and result-max
variables. Local variables affect only the elements which the will be pushed apart.
variables be defined and detailed in. Global variables affect In order to detect the state for termination, one more
all elements, and are detailed in the environment rather than variable, the settling boundary is defined [6]. Settling
in the elements. They are detailed in OPTION of MODEL in boundary is a percentage such that the termination criteria is
pull down menu of WITNESS. Since the variables
(parameters) are the tokens in genetic structures, the variable
type must also be defined throughout the whole system.
With the exception of these individual variables, some
parameters of elements may depend on one another. For
example, the number of machines of current station must be
the same as number of machines of next station. The number By combining the above developments and information,
of machines in both station is dependent. Another examples the optimization process can be established.
is that the length of conveyor may depends on number of 6 Result
machines output to it. In these cases, new variable set called In order to calculate the resource level required for the
association was defined in the system. These variables handle production line setup to obtain maximize output, the
the dependent parameters, while only one of the dependent GA-Aided Simulation Shell was utilized. The experimental
parameters occupies one token in structures. Other variables environment of an injection moulding shop producing casings
depending on this one will change according to the for audio products in order to determine resources required is
dependence and be exported to WCL file to alter the shown in Appendix 1. Totally 115 iterations equivalent to
simulation model. 6900 simulation cycles were performed to determine the
resource level. At the 102th iteration, optimum values have
In Genetic Algorithms Shell not been obtained as there
Some essential parameters have to be defined. They are
the population size, crossover rate and mutation rate. In our
1996-97 Annual Journal of IIE (HK) 33
were still fluctuation in results. However, at 115th iteration, the result
became stable and the parameters being obtained can be used for resource
planning and for line balancing.
7 Limitation of the Genetic Algorithms Shell
Based System
In the present stage, this GA's shell only processes single _goal, for
instance, maximize machine utilization or maximize output quantity. However,
multiple goals co-exist with multiple constraints are concerned. Although this
may limits the power of the system, it is still capable of observing the effect of
Genetic Algorithms on random events simulation. In addition, in using this
GA's shell, several qualification of users is required. Users must be familiar
with their own model and WITNESS modelling structure. One of the
qualification of the user is the ability to handle modeling in simulation
software, and in this case, WITNESS.
8 Conclusion
This paper illustrates a new concept of using GA's shell interfacing with
the simulation software to solve optimization problem which is beneficial to
line balancing and resource allocation problem. As GA's are applied on
random events, the suggestion from this GA -based decision support system is
logically an optimium solution of line balancing and resource allocation.
References
1. J.H. Holland, ADAPTION IN NATURAL AND
ARTIFICIAL SYSTEMS, U. of Michigan Press, 1975.
2. S.A. Kennedy, "Five Ways to a Smarter Genetic Algorithm",
A1 EXPERT,1993
3. L. Davis, HANDBOOK OF GENETIC ALGORITHMS, NY
Van Nostrand Reinhold, 1991.
4. L. Davis, "Genetic Algorithm Profiles in Advanced Technology for
Developers. Sewickley", Pa: High-Tech Communications, 1992.
5. L.D. Whitley, FOUNDATIONS OF GENETIC
ALGORITHMS, Wash. Axcelis Press, 1993.
6. J.J. Grefenstette, "Optimization of Control Parameters for Genetic
Algorithms", IEEE Trans. on Sys. Man & Cyb., 1986.
Appendix 1- Experimental Environment
Abstract 2 Review
Flexible manufacturing system (FMS) design and implementation There are several significant issues which are critical to
are lengthy and costly process. The lack of suitable reference models successful FMS development. These include the availability of
and connections between various development support tools further suitable system architecture; reusable software; and the use of
make the situation difficult. This paper addresses these problems by appropriate system design tools.
devising a generic flexible manufacturing cell (FMC) control
software architecture, and an integrated development environment Open system architecture
for FMC's. The control software architecture consists of a three-layer Open FMS development can be facilitated based on an open
hierarchy including the operation, networking and operating- system system reference model. An open system may be defined as
layer. This arrangement succeeds in separating the hardware and exhibiting the characteristics of high connectivity between
communication issues from the main FMC control applications. On subsystems, availability to a broad set of users, expandability of both
the other hand, the proposed development environment integrates hardware and software, and portability of software from one
several development tools to support system development from subsystem to another. Wright [ 1 ] adopts an anthropomorphic
design, evaluation, implementation, testing to deployment. Such a paradigm to derive the skeleton of the whole CIM operation
development environment streamlines FMS development process architecture. He suggests the use of de facto standards for languages,
and saves both time and achieves better FMS design. operating systems, and computer hardware to serve as a basis for
open system development, which results in higher feasibility of use
1 Introduction and in turn acceptability. Levin et al [2] suggests a centralised control
The development of FMS is usually treated as an one-off software structure with emphasis on software modularity, and
process. Such a system needs to be tailor-made to satisfy various accounting of real-time requirements in programming. The National
objectives according to specific corporate, market and Bureau of Standards (NBS) proposed the concept of virtual
environmental requirements. The lack of suitable reference models manufacturing cells and automated manufacturing research facility
and development tools forces development work to start from [3]. The system is optimised dynamically by integrating planning and
scratch. Moreover, the multi-objective nature of the problem scheduling functions together.
encumbers tremendous design work. Added to this, the
implementation of hardware and software components is both Software reuse methodologies
tedious and error prone. As a result, the development of FMS's is Nof [4] identifies the advantages and limitations of using
usually lengthy and costly. object -oriented methodology in manufacturing. He notices that the
On the other hand, many installed systems have not quite natural and intuitive correspondence, inherent structure, modularity,
achieved the goals set out by the FMS community. The usual discipline, adaptability and flexibility are beneficial to manufacturing
problems of those systems include low production rate and inability applications. However the complexity in handling distributed
to respond to product and volume changes. In other cases it is found communications and operations, the difficulties in co-ordinating
that such systems are difficult to maintain and to modify or upgrade. remote activities, the potential high run-time cost, and a lack of
The behaviour and structure of FMS's epitomise a complex system functional support and standard libraries limit its use. The reusable
that their design and implementation call for expert design software concept is also studied by Smith [5]. Important reusable
knowledge and techniques, and development tools especially for softwa re concept including the use of scaleable architecture,
control software development. A plethora of these techniques and automatic generation of control code and object -oriented design are
tools have been available. Unfortunately no effort has been made to identified. They are achieved through the use of hierarchical control
select the appropriate tools and to integrate them into an FMS skeleton, context free grammar modelling approach and
development environment to boost the development productivity object -oriented modelling of physical entities, respectively.
and quality of the FMS.
This paper addresses the issues of using a generic FMC control System design methodologies
architecture as a reference model for software development. Talavage [6] identifies variables which affect system
Additionally, a set of tools are provided under an integrated performance. These variables include process plans, materials,
development environment (IDE) to facilitate the design and configuration and operation control logic. By using simulation,
implementation of FMC's. The control software has a generic knowledge-based system technique and problem decomposition
structure which can be adopted to suit a wide range of system method, an optimal solution as a combination of the aforesaid
requirements. The control functions provided are comprehensive variables is found automatically. The problem decomposition method
and the control software can be easily maintained and modified. On relates practical problems to such variables. The application of
the other hand, the 1DE can significantly facilitate the design simulation depicts system performance with respect to change of
process, avoid design errors, reduce the need of expertise and variables. Chaar [7] reviewed different approaches and tools to
shorten the development horizon. developing manufacturing and realtime software. All these
approaches stress the importance of formal models and specification
languages in capturing functional and timing requirements of
real-time embedded
3 A Ge neric FMC Controller Architecture The networking level acts as a linkage between the operation
and operating system levels. It provides communication facilities for
Control hierarchy software entities at these two layers. The networking level is further
In this discussion, a two-level cell- station control structure is decomposed into two sub- levels: the (FMC) process and task level.
adopted. A station here is defined as processing, storage, transferring, The task level is designed to interface to particular operating system
transporting or a combination of these capabilities. It employs a and communication protocols chosen. It provides a comprehensive
centralised control hierarchy where a cell supervisor is used to set of communication functions which are useful for general
co-ordinate the tasks of all components in the cell. The major networking software development. By providing standard routines,
scheduling decisions concerning resource allocation, logistics and communication between local or remote software modules can be
contingency planning are generated by the FMC supervisor. Stations treated in the same manner. Organising the networking level
in the cell, however, have a high degree of local intelligence modules in this way achieves a clean logical structure of the
especially in monitoring their own operations. In the proposed software/hardware entities. Thus a conceptually homogeneous
system, centralised control is preferred to decentralised control environment is provided for software development, where
architecture. The former has a number of advantages inc luding the information concerning the location of each software module,
correspondence to corporate control hierarchy, simpler cell communication mechanism used between different computing
components co-ordination activities, lower communication devices and for different types of information can be mask ed. On the
requirement, easy adoption of different planning and scheduling other hand, the process level provides functionality which supports
methods, and more capable of generating optimal plans and FMC applications. Dedicated functions are built up on functionality
schedules in real time.
Abstract the real output per worker in the Hong Kong service sectors between
Hong Kong has gone through a transition from manufacturing 1986 and 1993 increased at an average rate of only 2.3% per annum
dominance to service dominance over the last 15 years. Traditionally compared to 8% for manufacturing. This clearly indicates an area
productivity is well defined and managed under the manufacturing where industrial engineers can contribute a quantum improvement
scenario. As the service industry continues to grow, there is a via innovative and viable ideas.
genuine need to study productivity in the service environment. This paper describes a detailed study of the various aspects of
In this paper, the providers of productivity awareness, the types productivity including the service providers, types of productivity
of productivity measures, the basic factors underlying productivity, measures, basic factors underlying productivity, and productivity
and productivity improvement techniques are discussed first. There is improvement techniques. The quality factor is also considered in a
an analysis on the benefits of productivity improvement and ideas on practical sense. Finally it shows how a good system aim ing at
how to treat employees fairly as a result of productivity gain. Then improving service productivity can lead towards Quality Excellence
the consideration of quality together with service productivity and that provides a strong platform to make customers totally satisfied.
how it leads to Quality Excellence are discussed. Concepts of Some real examples are quoted with reference to the DHL Hong
`negative productivity' and `quality level' discount factors are Kong operations.
introduced. Finally the approach of a leading air express company in
service productivity and quality excellence is discussed. 2 Providers of Productivity Awareness in Hong
Kong
1 Introduction Due to the need to survive through fierce market competition,
Hong Kong has already become a city led by service industries. every business organisation would strive to improve their own
According to Government statistics, the importance of the service productivity, particularly for those international companies. Usually
sector in Hong Kong is very significant, both in terms of they set up operations improvement teams, promote cost reduction or
contribution to the Gross Domestic Product (GDP) and the impact business re-engineering programs to make these happen.
on employment. The rapid transition of the Hong Kong economy In addition to these people, there are also some outside parties
from manufacturing to service dominance over the last 15 years is who are destined in promoting productivity of the Hong Kong
phenomenal. The opening up of China since 1980 to foreign industries.
investment was the main catalyst of these changes. This has allowed The first one to be mentioned is the Hong Kong Productivity
the Hong Kong manufacturers, who have been increasingly Council (HKPC). HKPC is a government subvented independent
squeezed by the huge rising costs of land and labour, to relocate body aiming at helping the industries in raising productivity and
their production facilities across the border into the Pearl River using more efficient methods. Their mission is to promote
Delta region of China. In the mean time, Hong Kong has gradually productivity excellence through professional services. HKPC
built a good infrastructure such as telecommunication networks, provides a wide range of services such as consulting service, product
banking systems, airport and sea container port facilities. These development, training and technology transfer to clients across the
factors, together with other favourable conditions e.g. the industrial sectors in Hong Kong. In a briefing workshop held in early
availability of high calibre people, free movement of money and 1996, HKPC announced that they would focus at offerring support to
goods in and out of the territory, have formed a strong base to Hong Kong trading firms and service organisations for productivity
support the continued growth of service industries in Hong Kong in improvement [2].
the future. There is a regional organisation by the name of Asian
There are three main core groups of services that have led the Productivity Organisation (APO). Their role is to promote the
growth of the Hong Kong economy, namely, transportation, productivity of the industries in the various Asia Pacific countries.
financial service, and communications [ 1 ]. Air express is one of They hold conferences around the region to exchange concepts,
the key sectors under the transportation field. It is forecast to grow methodologies and technological applications among fellow member
even faster in the years to come because courier delivery service countries on a regular basis.
can help customers squeeze time out of their supply chain cycle. As Another organisation is the Hong Kong Institution of Engineer s
a result, companies can respond to the customers' needs faster and (HKIE). Under HKIE, the Manufacturing and Industrial Engineering
push products out to the market earlier. Hence they become more Division (MIE) is heavily devoted to the upgrading of engineering
competitive than their rivals. At the same time, DHL and the other professionalism that covers productivity, quality, methods, process
major players all wish to capture a bigger market share of the air re-engineering, and manufacturing technology. A recent development
express business. Obviously, productivity improvement is an area to is that this MIE division is extending the engineers' formal training to
look into first. include service industry in addition to the manufacturing ones before.
`Service Productivit y' is a less familiar term than its counterpart This can attract more engineering graduates to join the service
of manufacturing productivity. According to KPMG Management industry and help enhancing the competitiveness via
Consultants' report in December 1995 [ 1 ],
References
[l] KPMG Management Consultants Ltd., report paper on
"Moving Forward by Adding Value : A study of Productivity
in Hong Kong's Trading Sectors", 1995.
OUTBOARD MARINE CORPORATION ASIA MR CHOI YUET SHING, ERIC MR CHUN PING FAI
O&M PROJECT EXECUTIVE
LTD HONGKONG TELECOM ENGINEERING DIRECTOR
MR CHEUNG CHUNG HEI MR CHOI YUN WAI SCHICK LTD
PRODUCTION MANAGER
MR. CHEUNG HINKEI GUONGZHOU LAM SOON FOOD MR CHUNGCHUNHO
PRODUCTS LTD PRODUCTION ENGINEER
MR CHEUNG KA HO
SYLVA INDUSTRIES LIMITED
MR CHEUNG KA HO ANTHONY MR CHONG MAN LUNG
MANAGING DIRECTOR MR CHUNG FUK LUN, ALLEN
MR CHEUNG KIN WAH TELEAN TECHNOLOGY LTD
CHIEF MANAGEMENT SERVICES OFFICER
LECTURER MR CHONG TSZ KIT
MR CHUNGHEIYIN
MR CHEUNG KING LOI
MR CHOW CHI WAI MR CHUNG HON MAN
AUDITOR BUSINESS DEVELOPMENT MANAGER
INCHCAPE TESTING SERVICES HK LTD MR CHUNG KAYEE
HONG KONG QUALITY ASSURANCE
MR CHUNG HI TUNG
AGENCY MR CHOW PUI YIN
PRODUCTION ENGINEER DEMONSTRATOR I
MRCHEUNGKW OKLEUNG
MR CHOW SAI LUN, FERGUS CITY UNIVERSITY OF HONG KONG
MR CHEUNG LOK MAN
MS. CHEUNG SAU KAN MR CHUNG KIM H.
MR CHOW SAU WAI DATA QUALITY AUDITOR
MR CHEUNG SHUN CHAN PRODUCT ENGINEER TNT EXPRESS WORLDWIDE CO LTD
MR LAIKWOKMING MR LAMYATCHING
MR LAITAKKIN
MR LAU CHI WING MR LEE SAI HIT, IVAN HK INTERNATIONAL TERMINALS LTD
MS.LEENGARYEE
MR NG CHI KEI
MAXTOR (HONG KONG) LTD MR TSE WOON KWAN PHILIPS HONG KONG LTD - LCD FACTORY