Global Perspective for Competitive Enterprise, Economy and Ecology

Shuo-Yan Chou • Amy Trappey Jerzy Pokojski • Shana Smith Editors

Global Perspective for Competitive Enterprise, Economy and Ecology

Proceedings of the 16th ISPE International Conference on Concurrent Engineering

123

Editors Shuo-Yan Chou, PhD Department of Industrial Management National Taiwan University of Science and Technology (NTUST) 43 Keelung Road, Section 4 Taipei 106 Taiwan R.O.C. sychou@im.ntust.edu.tw

Amy Trappey, PhD Department of Industrial Engineering and Management National Taipei University of Technology No.1, Section 3, Chung-Hsiao E. Road Taipei 106 Taiwan R.O.C

Jerzy Pokojski, PhD Inst. Machine Design Fundamentals Warsaw University of Technology Narbutta 84 02-524 Warszawa Poland

Shana Smith, PhD Department of Mechanical Engineering National Taiwan University 1 Roosevelt Road, Section 4 Taipei 106 Taiwan R.O.C.

ISBN 978-1-84882-761-5 e-ISBN 978-1-84882-762-2 DOI 10.1007/978-1-84882-762-2 Springer Dordrecht Heidelberg London New York British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Control Number: 2009929381 © Springer-Verlag London Limited 2009 Apart from any fair dealing for the purposes of research or private study, or criticism or review, aspermitted under the Copyright, Designs and Patents Act 1988, this publication may only bereproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms of licencesissued by the Copyright Licensing Agency. Enquiries concerning reproduction outside those termsshould be sent to the publishers. The use of registered names, trademarks, etc. in this publication does not imply, even in the absence ofa specific statement, that such names are exempt from the relevant laws and regulations and thereforefree for general use. The publisher makes no representation, express or implied, with regard to the accuracy of theinformation contained in this book and cannot accept any legal responsibility or liability for any errorsor omissions that may be made. Cover design: eStudioCalamar, Figueres/Berlin Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Preface

The proceedings contain papers accepted for the 16th ISPE International Conference on Concurrent Engineering, held in the vibrant city of Taipei, Taiwan, from July 20 to 24, 2009. The conference is a sequel of the conference series launched by the International Society for Productivity Enhancement and has constituted an important forum for international scientific exchange on Concurrent Engineering (CE).

CE appeared in the 80’s as a concept of parallel performing engineering design activities and integrating all related processes. This concept is based on general assumption that different components of product life cycle should be considered together and relatively early in the development process. The main goal of CE is to make processes more efficient and more resistant to errors. Substantive advantages can be achieved by adopting CE strategies and methodologies.

The last twenty years brought many changes in organization of product design and manufacturing. Engineers’ professions received narrower specializations. Engineers became present on global market. Sometimes firms create alliances. Engineers work in firms from suppliers to final producers. Engineers cooperate and collaborate cross the border of countries. They need to use different methods and tools supporting their engineering and development activities. As a result, CE has been further expanded to support many aspects of product development. Meantime the whole CE approach has got different forms and names and has become omnipresent. Industrial presence of CE differs in particular cases, from well-established corporation implementations to little small firm applications.

From the beginning the role of information systems in Concurrent Engineering was treated as an indispensible facilitation technology. First methodologies and tools were concentrated on offering possibility to contact people and processes, to make available right information and knowledge at the right time. The presence of computer tools in CE is treated as a standard and covers a spectrum of activities. The functions are no longer limited to passively managing data and processes; they have become intelligent and proactively assisting design and development activities.

If one looks now at what is going on and at how many different issues are important in design, manufacturing, supply, distribution, etc., it is evident why Concurrent Engineering context is so rich and so complicated, why we have so many CE specializations and why CE2009 Conference’s main topic is the following: Global Perspective for Competitive Enterprise, Economy and Ecology.

vi Preface

The plurality of CE specializations mentioned above was transformed on the following plurality of CE2009 Conference tracks: Systems Engineering, Advanced Manufacture, Product Design, Design for Sustainability, Knowledge Engineering, SCM, Collaborative Engineering, Web Technologies, Service Solutions. Apart of the enumerated tracks the conference has also seven special sessions: Special Session in RFID, in Collaborative Product Development, in Multi-disciplinary Design and Optimization, in Design Knowledge Utilization, in Competitive Supply Chain Performance, in Value Engineering, in Competitive Design.

The proceedings contain 84 papers by authors from 14 countries. If we concluded that they belong to different tracks and special sessions then we see how multi-perspective the content of this volume is. There are papers which are theoretic, conceptual and papers which have very strong industrial roots. There are also papers very detailed, made from a narrow view and very close to specific industrial case studies. We can also find papers which are based on real processes but which operate on abstractive models and which offer a bridge between an industrial reality and an academic research. This heterogeneous nature of Concurrent Engineering brings together diverse and significant contribution to product design and development, which is also what the proceedings intend to offer.

Concurrent Engineering doesn’t develop equally in each direction. The way of development depends on many factors. We think that the content of this volume reflects what is actual, noticeable and the issues’ variety in the present stage of CE methods and phenomena. As a consequence of this fact careful readers can build their own view of present problems and methods of CE.

Shuo-Yan Chou General Chair, CE 2009 National Taiwan University of Science and Technology, Taiwan

Amy Trappey General Co-Chair, CE 2009 National Taipei University of Technology, Taiwan

Jerzy Pokojski Program Chair, CE 2009 Warsaw University of Technology, Poland

Shana Smith Local Chair, CE 2009 National Taiwan University, Taiwan

Program Committee

General Chair: Shuo-Yan Chou, National Taiwan University of Science and Technology, Taiwan

General Co-chair:Amy Trappey, National Taipei University of Technology, Taiwan

Program Chair:Jerzy Pokojski, Warsaw University of Technology, Poland

Local Chair:Shana Smith, National Taiwan University, Taiwan

General Secretariat:Yi-Kuei Lin, National Taiwan University of Science and Technology, Taiwan

Scientific Chair:Chih-Hsing Chu, National Tsinghua University, Taiwan

Financial Chair:Kung-Jeng Wang, National Taiwan University of Science and Technology, Taiwan

Publicity Chair:Shih-Wei Lin, Chang Gung University, Taiwan

Logistics Chair:Chih-Hsien Chen, Fo Guang University, Taiwan

viii Program Committee

ISPE Technical Committee

Ahmed Al-Ashaab, Cranfield University, UK

Amy Trappey, National Taipei University of Technology, Taiwan

Geilson Loureiro, International Society for Productivity Enhancement, Brazil

Jerzy Pokojski, Warsaw University of Technology, Poland

John Cha, Beijing Jiaotong University, China

M.J.L. van Tooren, Delft University of Technology, Netherlands

Mike Sobolewski, Texas Tech University, Texas, USA

P.M. Wognum, Wageningen University, Netherlands

Parisa Ghodous, University of Lyon, France

Rajkumar Roy, Cranfield University, UK

Richardo Goncalves, Instituto de Desenvolvimento de Novas Tecnologias, Portugal

Ricky Curran, Delft University of Technology, Netherlands

Shuichi Fukuda, Stanford University, USA

International Program Committee (IPC)

Sung-Hoon Ahn, Seoul National University, Korea

Ahmed Al-Ashaab, Cranfield University, UK

John Cha, Beijing Jiaotong University, China

Kuei-Yuan Chan, National Cheng Kung University, Taiwan

Wei-Lun Chang, Tamkang University, Taiwan

Chih-Hsien Chen, Fo Guang University, Taiwan

Jahau Lewis Chen, National Cheng Kung University, Taiwan

Kai-Ying Chen, National Taipei University of Technology, Taiwan

Shuo-Yan Chou, National Taiwan University of Science and Technology, Taiwan

Chih-Hsing Chu, National Tsinghua University, Taiwan

Richard Curran, Delft University of Technology, Netherlands

Program Committee ix

Jerry Y.H. Fuh, National University of Singapore, Singapore

Shuichi Fukuda, Stanford University, USA

Liang Gao, Huazhong University of Science and Technology, China

Parisa Ghodous, University of Lyon, France

Richardo Goncalves, Instituto de Desenvolvimento de Novas Tecnologias, Portugal

Kazuo Hiekata, The University of Tokyo, Japan

George Huang, The University of Hong Kong, Hong Kong

Roger J. Jiao, Georgia Institute of Technology, Atlanta, USA

Ajay Joneja, Hongkong University of Science and Technology, Hong Kong

Weidong Li, Coventry University, UK

Yi-Kuei Lin, National Taiwan University of Science and Technology, Taiwan

Shih-Wei Lin, Chang Gung University, Taiwan

Geilson Loureiro, International Society for Productivity Enhancement, Brazil

Masato Inoue, The University of Electro-Communications, Japan

Takashi Niwa, The University of Tokyo, Japan

Jerzy Pokojski, Warsaw University of Technology, Poland

Rajkumar Roy, Cranfield University, UK

Shana Smith, National Taiwan University, Taiwan

Mike Sobolewski, Texas Tech University, USA

Qiang Su, Shanghai Jiaotong University

Chika Sugimoto, The University of Tokyo, Japan

Kenji Tanaka, The University of Tokyo, Japan

Tien-Lung Sun, Yuan Ze University, Taiwan

Kai Tang, Hong Kong University of Science and Technology, Hong Kong

Dunbing Tang, Nanjing University of Aerospace and Aeronautics, China

M.J.L. van Tooren, Delft University of Technology, Netherlands

Amy Trappey, National Taipei University of Technology, Taiwan

Kung-Jeng Wang, National Taiwan University of Science and Technology, Taiwan

x Program Committee

Charlie, C.L. Wang, The Chinese University of Hong Kong, Hong Kong

Gary Wang, Simon Fraser University, Canada

Yan Wang, University of Central Florida, USA

Nel Wognum, Wageningen University, Netherlands

Yong Zeng, Concordia University, Canada

Sponsors

ISPE : International Society for Productivity Enhancement

National Taiwan University of Science and Technology, Taiwan

National Taipei University of Technology, Taiwan

National Science Council, Taiwan

Chinese Institute of Industrial Engineers

SME : Society of Manufacturing Engineers Taipei Chapter

Contents

Advanced Manufacture

To Calculate the Sequence-Dependent Setup Time for a Single-Machine Problem with Uncertain Job Arriving Time .................................................................... 3 Ming-Hsien Yang, Shu-Hsing Chung, and Ching-Kuei Kao

A Two-Level Genetic Algorithm for Scheduling in Assembly Islands with Fixed-Position Layouts .................................................................................................. 17 Wei Qin and George Q. Huang

Exact Controllability for Dependent Siphons in S3PMR ............................................... 29 Yu-Ying Shih, Te-Chung Liu, Chui-Yu Chiu and D. Y. Chao

A New MIP Test for S3PGR2 ........................................................................................ 41 Yu-Ying Shih, D. Y. Chao and Chui-Yu Chiu

Simplifying Abrasive Waterjet Cutting Process for Rapid Manufacturing ................... 53 Nguyen Van Ut, Pisut Koomsap, and Viboon Tangwarodomnukun

Hybrid System Supporting Flexible Design of Flat Rolling Production Processes in Collaborative Environment ....................................................................... 61 Lukasz Rauch, Michal Front, Marek Bigaj, Lukasz Madej

Collaborative Engineering

Organization and Interoperation of Engineering Design Services in Service-Oriented Architecture ....................................................................................... 73 Nan Li, Jianzhong Cha and Wensheng Xu

xiv Contents

Taxonomy and Attributes of Business Collaborative Models: an Exploratory Study ..................................................................................................... 83 JrJung Lyu and Ping-Shun Chen

Applying Petri Net to Analyze a Multi-Agent System Feasibility- a Process Mining Approach ........................................................................................... 93 C. Ou-Yang ,Yeh-Chun Juan ,C.S. Li

The Impact on Global Logistics Integration System to Concurrent Collaborative Process .................................................................................................. 105 Yao Chin Lin and Ping Heng Tsai

Collaborative Product Development

Using DEA and GA Algorithm for Finding an Optimal Design Chain Partner Combination ......................................................................................... 117 Chih-Ling Chuang, Tzu-An Chiang, Z. H. Che and H. S. Wang

Adaptive Architecture for Collaborative Environment ............................................... 129 Youssef Roummieh and Parisa Ghodous

Conceptual Modeling of Design Chain Management towards Product Lifecycle Management ................................................................................................ 137 Wei Liu, Yong Zeng

Data Persistence in P2P Backup Systems .................................................................... 149 Rabih Naïm Tout, Parisa Ghodous, Aris Ouksel and Mihai Tanasoiu

Competitive Design

Using DEA Approach to Develop the Evaluation and Priority Ranking Methodology of NPD Projects .................................................................................... 159 Ling-Chen Hung, Tzu-An Chiang , Z. H. Che and H. S. Wang

An Exploration Study of Data-mining Driven PLM Implementation Approach ..................................................................................................................... 167 Chia-Ping Huang

Contents xv

Exploring the Links between Competitive Advantage and Enterprise Resource Planning (ERP) Upgrade Decision: A Case Study Approach ...................... 179 Celeste See-Pui Ng and Pei-Chann Chang

Development of a Cost Estimating Framework for Nanotechnology-Based Products ................................................................................. 193 Yuchun Xu, Rajkumar Roy, Gianluca Cassaro and Jeremy Ramsden

An Analogy Based Estimation Framework for Design Rework Efforts ...................... 203 Panumas Arundacahawat, Rajkumar Roy and Ahmed Al-Ashaab

Design for Sustainability

Greening Economy as a Key Solution to the Economic Crisis ................................... 215 Peter Yang and Injazz Chen

A Study on Evaluation of Environmental Effectiveness of Manufacturing Processes ............................................................................................. 223 Nozomu Mishima, Shinsuke Kondoh, Keijiro Masui, Masayoshi Yasuoka, Yuji Hotta and Koji Watari

Understanding the Waste Net: A Method for Waste Elimination Prioritization in Product Development .............................................................................................. 233 Marcus V. P. Pessôa , Warren Seering, Eric Rebentisch and Christoph Bauch

The Green Product Eco-design Approach and System Complying with Energy Using Products (EuP) Directive .................................................................................. 243 Amy J.C. Trappey, Meng-Yu Chen, David W. Hsiao and Gilbert Y.P. Lin

Developing an ISO 14048-Based EuP Integrated Service Platform for Evaluating Environment Impacts and Supporting Eco-Design in Taiwan .................. 255 Tzu-An Chiang, Hsing Hsu, Ping-Yu Chang, Hung-Chia Wei

Systematic Lean Techniques for Improving Honeycomb Bonding Process ................ 267 Chiun-Ming Liu and Min-Shu Chiang

Expanding Environmental Information Management: Meeting Future Requirements in the Electronics Industry .................................................................... 281 Eric Simmon , John Messina

xvi Contents

Rule-Based Recursive Selective Disassembly Sequence Planning for Green Design ......................................................................................................... 291 Shana Smith and Wei-Hsiang Chen

Design Knowledge Utilization

Investigation on Evaluation of Design Decision for Door-Shaped Structure by Using Systematic Knowledge Analysis .................................................................. 305 Zone-Ching Lin, Chen-Hsing Cheng

Knowledge Extraction System from Reports in Fabrication Workshops .................... 317 Kazuo Hiekata, Hiroyuki Yamato and Sho Tsujimoto

Knowledge based Sales Forecasting Model for Non-linear Trend Products ............... 327 Kenji Tanaka, Hideaki Miyata and Shoji Takechi

Knowledge Engineering

A Negotiation Strategy of Collaborative Maintenance Chain and Its Multi-Agent System Design and Development ........................................................... 337 Amy J.C. Trappey, Wei-Chun Ni and Chun-Yi Wu

Develop Non-Exhaustive Overlapping Partitioning Clustering for Patent Analysis Based on the Key Phrases Extracted Using Ontology Schema and Fuzzy Adaptive Resonance Theory ...................................................................... 349 Amy J.C. Trappey, Charles V. Trappey and Chun-Yi Wu

Performance Evaluation for an ERP System in Case of System Failures ................... 361 Shin-Guang Chen

Multi-Disciplinary Design and Optimization

An Approach Based on Rough Sets Theory to Design Space Exploration of Complex Systems .................................................................................................... 373 Xue Zheng Chu, Liang Gao, Mi Xiao, Wei Dong Li, Hao Bo Qiu

The Set-Based Multi-Objective Satisfactory Design for the Initial Design with Uncertainties in Collaborative Engineering......................................................... 381 Masato Inoue, Yoon-Eui Nahm and Haruo Ishikawa

Contents xvii

Minimizing Makespan for Server Testing with Limited Resource.............................. 389 Ping-Yu Chang, Ya-Ting Hsu, and Chin-An Cheng

Exchange of Heterogeneous Feature Data in Concurrent Engineering and Collaborative Design Environments ............................................................................ 395 Zhiyong Huang, Fazhi He, Xiaoxia Li, Xiantao Cai and Huajun Liu

An Ergonomic Assembly Workstation Design Using Axiomatic Design Theory ....... 403 Xiaoyong Wang, Dunbing Tang, Peihuang Lou

Heterogeneous Material-based Biomedical Product Development ............................. 413 W.D. Li, L. Gao, D.B. Tang and K. Popplewell

Research on Variable Parameter Set in Complex Multi domain Physical System and Its Repeatedly Simulating Arithmetic ...................................................... 423 Renwang Li, YiZhong Wu, Liang Gao, Zhansi Jiang, Zefei Zhu

Two Stage Ant Coordination Mechanisms for Sequencing Problem in a Mixed Model Assembly Line ................................................................................................. 431 Qiong Zhu, Jie Zhang

Product Design

A Study of Design by Customers: Areas of Application ............................................. 445 Risdiyono and Pisut Koomsap

Dual Lines Extraction for Identifying Single Line Drawing from Paper-Based Over Traced Freehand Sketch ..................................................................................... 455 Natthavika Chansri and Pisut Koomsap

A Formal Representation of Technical Systems.......................................................... 465 Baiquan Yan and Yong Zeng

Design Knowledge Assets Management with Visual Design Progress and Evaluation.................................................................................................................... 477 Gundong Francis Pahng and Mathew Wall

Product Redesign Using TRIZ and Contradictive Information from the Taguchi Method ..................................................................................................... 487 Ying-Ting Shen and Shana Smith

xviii Contents

Radio Frequency Identification (RFID)

Toward Full Coverage UHF RFID Services - An Implementation in Ubiquitous Exhibition Service ................................................................................ 501Tung-Hung Lu, Li-Dien Fu

Building a RFID Anti-Collision Environment for Conference and Exhibition Industry ...................................................................................................... 511 Min-Hsien Weng, Chih-Wei Chao, Kuo-Shu Luo, Li-Dien Fu, Tung-Hung Lu

Networking Dual-Pair-Tele-Paths for Logistic and Parking Structures with RFID Applications ...................................................................................................... 519 Li-Yen Hsu

Applying RFID to Picking Operation in Warehouses ................................................. 531 Kai-Ying Chen, Yu-Feng Hwang and Mu-Chen Chen

POC of RFID Application in Forest Sample Zone Investigation ................................ 541 Shiang-Shin Lin, Teh-Chang Wu, Jenn-Sheng Wu, Yi-Ping Huang, Ming-Hsiung Chang, Sheng-Wei Fan, Jiun-Jiue Liao

Cost Reduction of Public Transportation Systems with Information Visibility Enabled by RFID Technology ..................................................................................... 553 Shuo-Yan Chou, Yulia Ekawati

Competitive Supply Chain Performance

Modeling and Solving the Collaborative Supply Chain Planning Problems ............... 565 Y. T. Chen, Z. H. Che, Tzu-An Chiang, C. J. Chiang and Zhen-Guo Che

A Bi-objective Model for Concurrent Planning of Supplier Selection and Assembly Sequence Planning ...................................................................................... 573 Y. Y. Lin, Z. H. Che, Tzu-An Chiang, Zhen-Guo Che and C. J. Chiang

Automobile Manufacturing Logistic Service Management and Decision Support Using Classification and Clustering Methodologies ...................................... 581 Charles V. Trappey, Amy J.C. Trappey, Ashley Y.L. Huang, Gilbert Y.P. Lin

Lead Time Reduction by Extended MPS System in the Supply Chain ....................... 593 JrJung Lyu and Hwan-Yann Su

Contents xix

A Multi-Product EPQ Model with Discrete Delivery Order: a Langrangean Solution Approach ....................................................................................................... 601 Gede Agus Widyadana, Hui Ming Wee

A Case Study on Impact Factors of Retailing Implementing CPFR - A Fuzzy AHP analysis ............................................................................................................... 609Hsin-Pin Fu, Sheng-Wei Lin, and Chi-Ren Chen

Autonomous Capacity Planning by Negotiation against Demand Uncertainty ........... 621 Shih-Min Wang and Kung-Jeng Wang

A Negotiation-Based Approach to Supply Chain Planning and Scheduling Problems in a Fully Distributed Environment ............................................................. 633 K. Robert Lai and Bo-Ruei Kao

Environmental Transparency of Food Supply Chains - Current Status and Challenges ............................................................................................................ 645Nel Wognum, Harry Bremmers

Multi-Product Min-Cost Recycling Network Flow Problem....................................... 653 Chiao-Lin Deng, Chun-Mao Shao

Service Solutions

Applying RFM Model and K-Means Method in Customer Value Analysis of an Outfitter .............................................................................................................. 665 Hsin-Hung Wu, En-Chi Chang and Chiao-Fang Lo

An Investigation of Community Response to Urban Traffic Noise ............................. 673 Ghorbanali Mohammadi

A Market Segmentation System for Consumer Electronics Industry Using Particle Swarm Optimization and Honey Bee Mating Optimization ........................... 681 Chui-Yu Chiu, I-Ting Kuo and Po-Chia Chen

Why the Big Three Decline Despite Their Lean Management - A Study Based on the Theory of Constraints ....................................................................................... 691 Simon Wu and H. M. Wee

xx Contents

The Business Data Integrity Risk Management Model: A Benchmark Data Center Case of IT Service Firm ................................................................................... 701 M. K. Chen, Shih-Ching Wang

The Key Dimensions for Information Service Industry in Entering Global Market: a Fuzzy-Delphi & AHP Approach ................................................................. 713 M. K. Chen, Shih-Ching Wang

Problem-Based Construction of Engineering Curricula for Multidisciplinary and Concurrent Engineering Practice .......................................................................... 725Gloria Lucía Giraldo and German Urrego-Giraldo

Competences Supported on Thematic Contents for Evaluation of Curricula Aiming to Concurrent Engineering ............................................................................. 735Gloria Lucía Giraldo and German Urrego-Giraldo

Predicting the Yield Rate of DRAM Modules by Support Vector Regression ............ 747 Shih-Wei Lin and Shih-Chieh Chen

Reflective Concurrent Engineering – 3rd Generation CE ............................................ 757 Shuichi Fukuda

The Study of Autonomous Negotiation System Based on Auction Enabled Intelligent Agent – Using Parking Tower Asset Maintenance as Case Example ........ 769 Yu-Lin Liu, David W. Hsiao and Amy J.C. Trappey

Value Engineering

KBE and Manufacturing Constraints Management ..................................................... 783 Richard Curran, Wim J.C. Verhagen, Ton H. van der Laan, MJT van Torren

Manufacturing Cost Contingency Management: Part a) Methodology Development .............................................................................. 793 Richard Curran, Marc Gilmour, C. McAlleenan, P. Kelly

Manufacturing Cost Contingency Management: Part b) Application and Validation .............................................................................. 803Richard Curran, Marc Gilmour, C. McAlleenan, P. Kelly

Systems Engineering Methodology for Concurrent Engineering Education ............... 813 Richard Curran, Michel van Tooren and Liza van Dijk

Contents xxi

Creating Value by Measuring Collaboration Alignment of Strategic Business Processes ...................................................................................................... 825 Frank van der Zwan, Sicco Santema, Richard Curran

Drivers of Customer Satisfaction in a Project-Oriented, Business-to-Business Market Environment: an Empirical Study ................................................................... 833 Wim J.C. Verhagen, Wouter W.A. Beelaerts van Blokland, Richard Curran

Web Technologies

Development of a Web-Based Mass Customization Platform for Bicycle Customization Services ............................................................................................... 847 Tung-Hung Lu, Amy J.C. Trappey

A Manufacturing Grid Architecture Based on Jini and SORCER ............................... 855 Wensheng Xu, Jianzhong Cha

Minding the Gap Between First and Continued Usage: an Empirical Study of the Implementation of a Corporate e-Learning English-Language Program at a Financial Firm in Taiwan ...................................................................................... 865 Tainyi (Ted) Luor, Hsi-Peng Lu, Robert E. Johanson, Ling-Ling Wu

Web-Based Mechanism Design of 3C Plastic Parts with Knowledge Management ............................................................................................. 877 Wen-Ren Jong, Chun-Cheng Lin, Yu-Hong Ting, Chun-Hsien Wu, Tai-Chih Li

WIW - A Web-Based Information System for Profile of Wind .................................. 889 Wu Xiao Bing, Adans Iraheta Marroquín, and Moacyr Fauth da Silva Jr.

Author Index................................................................................................................ 899

Advance Manufacture

To Calculate the Sequence-Dependent Setup Time for a Single-Machine Problem with Uncertain Job Arriving Time

Ming-Hsien Yanga,1, Shu-Hsing Chungb, and Ching-Kuei Kaob

aDepartment of Business Management, National United University, Miao-Li, Taiwan, ROC. bDepartment of Industrial Engineering and Management, National Chiao Tung University, Hsinchu, Taiwan, ROC.

Abstract: Consider a finite-capacity single machine responsible for processing several product types of jobs, in which the inter-arrival time of jobs of specific type is independently and exponentially distributed to reflect the uncertainties in market demand. Since the uncertainties of the time and the type of customer demand make the machine setup decision very complicated, the development of analytic models to calculate the expected sequence-dependent setup time under FIFO and FSR rules contributes to the quick evaluation of capacity waste due to changing machine setting. Analytic model also shows the saving of machine utilization rate by changing the considered dispatching rule from FIFO to FSR. (1E-mail: ymh@nuu.edu.tw (M. H. Yang))

Keywords: analytic models, expected sequence dependent setup time, Poisson process jobs arriving, capacity reduction, change the setting of machine

1 Introduction

We consider a manufacturing system designed for producing several types of products according to customer order, machine setup to switch from the current setting to a different one usually is necessary and can not be regarded as a part of job processing time. When a job of specific product type, indicating a customer order, arrives at the system, it will enter the queue line and waits a period of time according to the dispatching or scheduling rule, and then starts the job processing on machine. In this paper, the first-in first-out (FIFO) rule and the family-based scheduling rule (FSR) are considered, in which FSR [1, 2, 3] can reduce the setups, comparing with FIFO. Before starting the processing of a new job, the machine should pause and proceeds setup if the previously completed job is different from the new job. Therefore, not only job processing but also machine setup consumes

1 Corresponding Author : Department of Business Management, National United University, Miao-Li, Taiwan, ROC; Tel: 886-(37)-381591; Fax: 886-(37)-332396; E-mail: ymh@nuu.edu.tw

4 M. H. Yang, S. H. Chung and C. K. Kao

machine capacity, which indicates that total setup time is related to the efficient utilizing of machine capacity.

There are two types of setups, (1) sequence independent and (2) sequence dependent, the second is a generality of the first and is considered in this paper. Besides, there are two levels of setup time problem [4], including the lower level concerning the role of setup in scheduling jobs [5, 6, 7], and the higher level concerning that in decision making at the production planning and control [8]. Missbauer [8] and Vieira et al. [6, 7] considered the dynamic behavior of job arriving in their researches. Missbauer [8] defined the probability of a setup is necessary for an arriving job complying with Poisson process underlying the FIFO and the FSR dispatching rules, however, did not address the setup probability for an arriving job of specific product type. Vieira et al. [6 ,7] assumed an independent Poisson arrival of a job with product type j and arrival rate j and defined the probability of requiring a setup in a time interval. However, their equations simplified the setup probability to constantly be (1 1/J) without considering the effect of the product type of an arriving job.

To analyze the accuracy of proposed models, an experimental design with various arriving conditions among several types of product, which corresponds to various resource utilization rates, is conducted.

2 Development of the probabilistic model for setups

The number of setups and the sequence dependent setup time are observed for a period of time RT, which begins initially at time 0. Let Nj(t) be the number of arriving jobs of product type j by time t. That is, Nj(t) is the random variable to occur in the fixed interval (0, RT] and is a Poisson process with arrival rate j, and there would be nj independent jobs in the time interval (0, RT] for the product type j, where nj j×RT, j = 1, 2, …, J, and J is the number of product types. The total arrival rate is equal to the sum of j, 1

Jj j .

2.1 Development of the probabilistic model of setups under FIFO

A setup is necessary given the condition that if a job arrives in the time interval (0, RT] encountering a previously completed job of different product type, in which there are no jobs or at lease one job in the system. Considering the ith arriving job of product type j, then the waiting time for this job (Tij) is a gamma variable with parameters i and j. Thus, the probability for this job arriving in the time interval (0, RT] is shown as Equation (1), and the probability for this job arriving in time interval (0, RT] with a setup under FIFO (Ps,ij,FIFO) is shown as Equation (2).

1

0Pr j ij

iRT tj i

ij ij ijT RT t e dti

(1)

, , 0, ,1

Pr 1 1j js ij FIFO ij FIFO n FIFO

n

P T RT p p (2)

To Calculate the Sequence-Dependent Setup Time for a Single-Machine Problem 5

where the probability of an arriving product type j job requiring a setup is (1 j/ ),p0,FIFO and pn,FIFO represent the probabilities that there are no jobs and there aren 1 jobs in the system under FIFO respectively.

According to the probability Ps,ij,FIFO, this arriving job faces two possible cases, in which setup occurs with the probability Ps,ij,FIFO and no setup occurs with the probability (1 Ps,ij,FIFO). Thus, the expected number of setups for this job is E[NSij,FIFO] Ps,ij,FIFO. If there would arrive nj independent product type j jobs in the time interval (0, RT], then the expected number of setups for product type j is calculated as 1, ,[ ] [ ]jn

ij FIFO ij FIFOE NS E NS , where j=1, 2, …, J.Equation (3) shows the expected setup time under FIFO. Let sjr be the setup

time for the job of product type j, in which the previously processed job belongs to product type r. The first part indicates this job does not arrive in the time interval (0, RT], then the setup time should be zero with the probability Pr[Tij>RT]. The second part indicates this job arrived in the time interval (0, RT] and in the same product type j as previously processed job on machine, the setup time would be sjj

0 and the probability would be Pr[Tij RT]( j/ ). The third part indicates a different type r job arrived in the time interval (0, RT] and the setup time would be sjr, then the probability would be Ps,ij,FIFO( r/ c), where 1,

c Jr r j r .

, , ,1

, ,1

Pr 0 Pr

Jr

ij FIFO ij ij j jj s ij FIFO jrcrr j

Jr

s ij FIFO jrcrr j

E S T RT T RT s P s

P s (3)

Suppose that there are nj independent product type j jobs arrive in the time interval (0, RT], then the expected mean setup time for product type j jobs arriving in the time interval (0, RT] is defined as 1, ,[ ] [ ]jn

ij FIFO ij FIFO jE S E S n , where j=1,2, …, J. Finally, the expected mean setup time of jobs arriving in the time interval (0, RT] is derived as 1 11 ,[ ] [ ]jnJ J

j jiFIFO ij FIFO jE S E S n .

2.2 Development of the probabilistic model of setups under the FSR

Comparing with FIFO, the difference in the probability of setups under FSR occurring at the condition there are n 1 jobs in the system. Suppose that there are n 1 jobs in the system, a setup is needed given the condition if a job of type jarrives in the time interval (0, RT] encountering none job in queue belonging to type j then the probability would be (1 j/ )n. Therefore, the probability that a setup is necessary for the ith job of product type j under FSR is shown as Equation (4).

, , 0, ,1

Pr 1 1n

j js ij FSR ij FSR n FSR

n

P T RT p p (4)

where p0,FSR and pn,FSR are the probabilities that there are no job and n 1 jobs in the system under FSR respectively.

6 M. H. Yang, S. H. Chung and C. K. Kao

Then the expected number of setups for the ith job of product type j is shown as E[NSij,FSR] Ps,ij,FSR. Suppose that there would arrive nj independent product type jjobs in the time interval (0, RT], which indicates the expected number of setups for product type j is computed as 1, ,[ ] [ ]jn

ij FSR ij FSRE NS E NS , where j=1, 2, …, J.Furthermore, the expected setup time under FSR is computed according to the

same idea of Equation (3), to replace Ps,ij,FIFO with Ps,ij,FSR, then the calculation of the expected setup time for the ith job of product type j in the time interval (0, RT]under FSR is shown as 1,, , ,[ ] ( )cJ

r r jij FSR s ij FSR r jrE S P s . Then the expected mean setup time for product type j jobs arriving in the time interval (0, RT] is computed as 1, ,[ ] [ ]jn

ij FSR ij FSR jE S E S n , where j=1, 2, …, J. The expected mean setup time of jobs arriving in the time interval (0, RT] is derived as

1 11 ,[ ] [ ]jnJ Jj jiFSR ij FSR jE S E S n .

3 Accuracy analysis of the probabilistic model of setups under FIFO and FSR

Suppose that there is a sequence of jobs that has been dispatched by FIFO and FSR respectively. We are interest to the achieved magnitude of reduction in capacity utilization rate by replacing the FIFO with FSR. For single machine system, the machine utilization rates under FIFO and FSR are shown as FIFO E[STFIFO] and

FSR E[STFSR], where E[STFIFO] and E[STFSR] are the expected mean service time of jobs under FIFO and FSR.

Suppose that the service time of jobs equals its processing time plus its setup time, and then the expected mean service time of jobs under FIFO is calculated according to Equation (3). First, the service time of the ith product type j job would be zero with the probability Pr[Tij>RT]. Second, the service time of the ith product type j job would be equal to its processing time with the probability Pr[Tij�RT]( j/ ). Third, the service time of the ith product type j job would be equal to its processing time plus its setup time with Ps,ij,FIFO( r/ c), where r 1, 2, …, Jand r j. Let STij,FIFO be the random variable of the service time of the ith job of product type j under FIFO. Then, the expected mean service time of jobs under FIFO is derived as 1 11 ,[ ] [ ]jnJ J

j jiFIFO ij FIFO jE S E S n and is given as Equation (5). To replace E[SFIFO] with E[SFSR], then the expected mean service time of jobs under FSR (E[STFSR]) is shown as Equation (6).

1

1 1 1

PrjnJ J

FIFO j ij j FIFOj j i

E ST n T RT pt E S (5)

1

1 1 1

PrjnJ J

FSR j ij j FSRj j i

E ST n T RT pt E S (6)

Substituting Equation (5) into Equation (6), then E[STFSR] is rewritten as E[STFSR]E[STFIFO] E[SFIFO] E[SFSR].

To Calculate the Sequence-Dependent Setup Time for a Single-Machine Problem 7

Comparing the effects of the setup time and the machine utilization between FIFO and FSR

Because the probability p0,FSR and pn,FSR are equal approximately to (1 FSR)and (1 FSR)( FSR)n. Then, E[SFSR] is reformulated as Equation (7). The part of Equation (7) is shown as Equation (8). It is seen that (1+( FSR/(1 FSR))( j/ )) 1with 0 FSR 1 and j > 0 for all j, which implies that E[SFIFO] E[SFSR], where the setups is reduced by FSR if the probability of the number of jobs in queue occurs. It is noticed that E[STFSR] E[STFIFO] E[SFIFO] E[SFSR] E[STFIFO]. Thus it is seen that FSR E[STFSR] FIFO E[STFIFO], which implies that the machine utilization under FSR is small or equal to the machine utilization under FIFO.

1 1

, ,1 1 1 1

1

, ,1 1 1 1

1 1 11

j

j

nJ J JjFSRr

FSR j s ij FIFO jr FSRcj j i r FSR

r j

nJ J Jr

j s ij FIFO jr FIFOcj j i r

r j

E S n P s

n P s E S

(7)

,1

,1

1, if 0 with 0, 1

11, if 0< 1 with 0,

FSR n FSR jj nFSR

FSRFSR n FSR j

n

p j

p j (8)

Relationship of the machine utilization rate between FIFO and FSR

The reduced range of the machine utilization rate by changing FIFO into the FSR is defined as FIFO FSR E[STFIFO] E[STFSR]. Thus, the machine utilization by changing FIFO into FSR is defined as FIFO FSR FIFO , where

is obtained by solving the simultaneous equations E[STFSR] E[STFIFO]E[SFIFO] E[SFSR] and E[STFSR] FSR/ . Substituting E[STFSR] FSR/ into the equation E[STFSR] E[STFIFO] E[SFIFO] E[SFSR], then this equation is rewritten as Equation (9).

1

1 1 1

Prj

FSRFSR FSR

nJ JFSR

j ij j FSRj j i

f E ST

n T RT pt E S (9)

where E[SFSR] is given by Equation (7). Because f( FSR) is differentiable, the Newton’s method is used to solve the nonlinear equation f( FSR) 0.

According to f( FSR) and its derivative with respect to FSR, we begin with a first guess 00 1FSR . A approximate solution 1

FSR can be obtained by calculating 0 0 0( ) ( )FSR FSR FSRf f , in which should be an even better approximation to the

solution of f( FSR) 0. Once we have 1FSR , we can repeat the process to obtain

2FSR . After n steps, if we have an approximate solution n

FSR , then the next step is

8 M. H. Yang, S. H. Chung and C. K. Kao

to calculate 1 ( ) ( )n n n nFSR FSR FSR FSRf f . Notice that if the value for n

FSRbecome closer and closer to 1n

FSR . This means that we have found the approximate solution of f( FSR) 0 after n steps.

Given a specific total arrival rate ( *) and the vector of job processing time (PT) in Figure 1, a slope and a intercept are given by *1 and zero for a line

*[ ]E ST , where x-axis represents the machine utilization rate and y-axis represents the expected service time. Thus, the expected service time under FIFO is calculated by Equation (5) ( * ,[ ]FIFOE ST PT ) given * and PT, and then the machine utilization rate under FIFO ( * ,FIFO PT ) equals *

*,[ ]FIFOE ST PT . Next, the

expected service time under FSR (E[STFSR]) and the function f( FSR) 0 for various FSR are depicted in Figure 1 according to Equation (5) and Equation (9). It is seen

that the function f( FSR) 0 is E[STFSR] to shift down and the shift quantum is FSR/ .Then, a root of f( FSR) 0 ( FSR ) is found by using the Newton’s method. Substituting FSR into E[STFSR], thus * , ,[ ]

FSRFSRE ST PT is obtained given

FSR FSR . Thus, the machine utilization by changing FIFO into FSR is given by Equation (10).

* * * ** *

, , , , FSRFIFO FSR FIFO FSRE STPT PT PT (10)

Then, the total arrival rate is changed from * to ** and with the same vector of job processing time (PT), where **< *. Thus, the slope of a line **[ ]E ST is equal to **1 , where ** *1 1 . Repeating the above step, the machine utilization by changing FIFO into FSR is derived by Equation (11) given the condition that **.

** ** ** **** **

, , , , FSRFIFO FSR FIFO FSRE STPT PT PT (11)

To Calculate the Sequence-Dependent Setup Time for a Single-Machine Problem 9

* ,[ ]FIFOE ST

PT

[ ]FSRE ST

* ,( )FSRf PT

*FSR

0

*E ST

** ,[ ]FIFOE ST

PT

[ ]FSRE ST

** ,( )FSRf PT

**FSR

0

*E ST**E ST

( , ) FIFOFIFO FIFOf E ST PT PTPT

( , ) FSRFSR FSRf E ST PT PTPT

* ,FIFO PT

*FIFO FSR

* , ,[ ]

FSRFSRE ST

PT

** ,FIFO PT**FIFO FSR

** **, ,[ ]

FSRFSRE ST

PT

Figure 1. Relationship of the machine utilization rate between FIFO and FSR

Thus, the bold lines in Figure 1 are depicted, which are the relationships between the machine utilization rate and the expected service time for various total arrival rates under FIFO and FSR.

Moreover, is rewritten as Equation (12) according to Equation (7).

11

, ,1 1 1 1

1 11

j

FIFO FSR FIFO FSR

nJ J JjFSR r

FSR j s ij FIFO jrcj j i rFSR

r j

E ST E ST E S E S

n P s (12)

Thus, the derivative of with respect to FSR is written Equation (13). 1 1

, ,1 1 1 1

2

2

1 11

1 011

jnJ J JjFSRr

j s ij FIFO jrcj j i rFSR FSR

r j

j jFSR FSR

FSRFSR

d n P sd

(13)

where d /d FSR 0 with 0 FSR 1 and j > 0. Hence, if FSR1 FSR2, then ( FSR1) ( FSR2), where 0 FSR1 1 and 0 FSR2 1. By replacing FIFO

10 M. H. Yang, S. H. Chung and C. K. Kao

with FSR, the reduced range of the machine utilization rate is increasing when the utilization rate of machine is rasing, which implies that the machine utilization rate is saved by FSR especially at the condition of high workload on machine.

3.1 Experimental design

To evaluate the accuracy of the probabilistic model on calculating expected setups and setup time, a simulation model is built to simulate the inter-arrival time of jobs with exponential distribution, in which the next job selected from queue to be processed on machine follows FIFO or FSR rule. If the selected next job requires changing machine setup, the sequence-dependent setups is included.

The simulation is implemented according to the experimental design by varying three control parameters: run time (RT), total arrival rate ( ), and the coefficient of variation of the jobs arrival rate (CV). Three levels of run time (RT) are considered: 8, 16, and 24 hours. The vector of job arrival rate among eight product types is shown as [ 1 2 3 4 5 6 7 8], and then six levels of the total arrival rate ( ) are considered as a ( 8

1j ja ) jobs in 60 seconds, where a = 1.00, 0.95, 0.90, 0.85, 0.80, and 0.75.

The coefficient of variation of the jobs arrival rate is considered, where the coefficient of variation is reported as a percentage and calculated from the mean and the standard deviation of the arrival rate of jobs. Table 1 shows three vectors of the arrival rate among eight product types and their corresponding CV, in which the total arrival rate of these three vectors are all equal to 0.01. Thus, the coefficient of variation of these three vectors of the arrival rate is calculated as 0, 27.9753, and 53.7234, which implies that the dispersion of the jobs arrival rate is greater when the coefficient of variation is increasing.

Table 1. Three vectors of arrival rate among eight product types and their corresponding CV

Arrival rate vector

Product type CV(%)1 2 3 4 5 6 7 8

1 0.001250 0.001250 0.001250 0.001250 0.001250 0.001250 0.001250 0.001250 02 0.001342 0.001577 0.001804 0.000917 0.001145 0.001443 0.000817 0.000955 27.9753 3 0.001821 0.001255 0.000297 0.001721 0.000946 0.000363 0.001467 0.002130 53.7234

For each combination, the simulation model contains the vector of job processing time among eight product types, PT = [15 75 85 45 55 10 80 125], and the matrix of sequence-dependent setup time (ST), shown below, for switching the setting of machine in changing product types. “Second” is the unit of processing time and setup time. Note that the simulation results are collected for each combination in 8 hours after 10,000 independent simulation runs.

To Calculate the Sequence-Dependent Setup Time for a Single-Machine Problem 11

045751545604530450607560759045301504530907530604575045904515301530450453015156030459007560304530757560090156045604530150

ST

3.2 Sensitivity analysis of the probabilistic model

Figure 2 shows the expected mean setup time per job (E[SFIFO] and E[SFSR]) for varying the CV of the job arrival rate, the total arrival rate, and the run time under FIFO and FSR in the probabilistic model. Figure 3 shows the mean setup time per job by the simulation model. Comparing with the data in Figure 2 and Figure 3, the mean setup time per customer by the simulation model seems to be a bound, E[SFIFO] and E[SFSR] are close to this bound as the total arrival rate and the run time increase. Moreover, in Figure 2 and Figure 3, for CV=53.7234%, the setup time under FIFO and FSR by the probabilistic model and the simulation model have the lowest level of value. As the increasing of the coefficient of variation of job arrival rates, the arrival of job tends to concentrate on fewer product types, which are the types having possible reduction of setups.

3.3 Accuracy of the probabilistic model

To test whether there is significant differences of setup time between FIFO and FSR ( ST), the statistical paired t-test is used. This test is to validate or falsify the null hypothesis that the difference of the setup time between FIFO and FSR is small or equal to zero and the alternative hypothesis that this difference is greater than zero. Table 2 presents the results of the paired t-test. With the P-value list in Table 2, we conclude that the difference of the setup time between FIFO and FSR is larger than zero, which implies that FSR can be used to save the setup time.

Figure 4 plots the mean absolute error of expected setups under FIFO and FSR, in comparison with the data from simulation model, by varying the run time (RT),the coefficient of variation (CV) of the arrival rate of jobs, and the total arrival rate. The mean absolute error of setups under FIFO is smaller than that of FSR, which implies that the probabilistic model under FIFO can calculate the setups more accurately than that under FSR. The overall mean absolute errors under FIFO and FSR are 3.3632 times and 5.0404 times respectively. Moreover, for CV equal to 27.9753% having lowest absolute error of expected setups. When the total arrival rate is increasing, the absolute error increases correspondingly, that is, the larger absolute error occurring at higher level of the utilization rate of machine.

Figure 5 displays the mean absolute error percentage of expected setup time under FIFO and FSR, in comparison with the data from simulation model, by

12 M. H. Yang, S. H. Chung and C. K. Kao

varying the run time (RT), the coefficient of variation (CV) of the arrival rate of jobs, and the total arrival rate. The mean absolute error percentage of setup time for the model under FIFO is also smaller than that under FSR. The overall mean error percentage of setup time under FIFO is 6.4080% and that under FSR is 9.2299%. Comparing with FSR, the probabilistic model under FIFO can calculate the setup time more accurately. Furthermore, when the CV equals 27.9753%, there has the lowest absolute error percentage of expected setup time. While the total arrival rate is increasing, the absolute error percentage increases respectively, that is, the larger absolute error percentage occurring at higher level of the machine utilization rate.

Figure 6 shows the differences of the number of setups and the setup time by varying the CV of the job arrival rate, the total arrival rate and run time under FIFO and FSR by the probabilistic model and the simulation model. For the probabilistic model, the reduced ranges of the number of setups and the setup time, by replacing FIFO with FSR, increases as the raising of the total arrival rate, which are shown in Figure 6(b) and 6(d). However, for the simulation model, the reduced ranges of the number of setups and the setup time, by replacing FIFO with FSR, increases firstly and then decreases as the raising of the total arrival rate, which are shown in Figure 6(a) and 6(c). Furthermore, we can see that the reduced rage of the setup time by replacing FIFO with FSR has larger magnitude at CV=53.7234%. This means that the reduced rage of the setup time due to the replacing is increasing as the enlarging of the coefficient of variation of the job arrival rates.

4 Conclusions

This paper provides the probabilistic models to calculate the number of setups and the sequence-dependent setup time for a single-machine system facing uncertain job arrivals, that is, the inter-arrival time of jobs being exponentially distributed. Since the uncertainties of the time and the type of customer demand make the machine setup decision very complicated, the development of analytic models to calculate the expected sequence-dependent setup time under FIFO and FSR rules contributes to the quick evaluation of capacity waste which is due to the changing of machine setting among several product types.

The computational results show that the overall mean absolute errors of setups under FIFO and FSR are 3.3632 times and 5.0404 times respectively, and the overall mean absolute error percentages of setup time under FIFO and FSR are 6.4080% and 9.2299% respectively, which both are related to the coefficient of variation (CV) of the arrival rate of jobs and the total arrival rate. In general, the probabilistic models under FIFO rule can calculate the number of setups and the sequence-dependent setup time more accurately than those under FSR. However, comparing with FIFO, FSR can reduce setups and then leads to the saving of utilization rate of machine, especially at the condition of higher total arrival rate. Therefore, our models can, in some extent, accurately calculate the total setup time and efficiently evaluate the capacity waste coming from changing machine setting to respond to the uncertainties of job arrivals.