Discrete-Event Simulation and System Dynamics for Management Decision Making E-Book

• Herausgeber: John Wiley & Sons
• Kategorie: Fachliteratur
• Serie: Wiley Series in Operations Research and Management Science
• Sprache: Englisch

Seitenzahl: 658

Veröffentlichungsjahr: 2014

Beschreibung

In recent years, there has been a growing debate, particularly in the UK and Europe, over the merits of using discrete-event simulation (DES) and system dynamics (SD); there are now instances where both methodologies were employed on the same problem. This book details each method, comparing each in terms of both theory and their application to various problem situations. It also provides a seamless treatment of various topics--theory, philosophy, detailed mechanics, practical implementation--providing a systematic treatment of the methodologies of DES and SD, which previously have been treated separately.        

Leseprobe

CONTENTS

Cover

Wiley Series in Operations Research and Management Science

Title Page

Copyright

Dedication

Preface

List of Contributors

Chapter 1: Introduction

1.1 How this Book Came About

1.2 The Editors

1.3 Navigating the Book

References

Chapter 2: Discrete-Event Simulation: A Primer

2.1 Introduction

2.2 An Example of a Discrete-Event Simulation: Modelling a Hospital Theatres Process

2.3 The Technical Perspective: How DES Works

2.4 The Philosophical Perspective: The DES Worldview

2.5 Software for DES

2.6 Conclusion

References

Chapter 3: Systems Thinking and System Dynamics: A Primer

3.1 Introduction

3.2 Systems Thinking

3.3 System Dynamics

3.4 Some Further Important Issues in SD Modelling

3.5 Further Reading

References

Chapter 4: Combining Problem Structuring Methods with Simulation: The Philosophical and Practical Challenges

4.1 Introduction

4.2 What are Problem Structuring Methods?

4.3 Multiparadigm Multimethodology in Management Science

4.4 Relevant Projects and Case Studies

4.5 The Case Study: Evaluating Intermediate Care

4.6 Discussion

4.7 Conclusions

Acknowledgements

References

Chapter 5: Philosophical Positioning of Discrete-Event Simulation and System Dynamics as Management Science Tools for Process Systems: A Critical Realist Perspective

5.1 Introduction

5.2 Ontological and Epistemological Assumptions of CR

5.3 Process System Modelling with SD and DES through the Prism of CR Scientific Positioning

5.4 Process System Modelling with SD and DES: Trends in and Implications for MS

5.5 Summary and Conclusions

References

Chapter 6: Theoretical Comparison of Discrete-Event Simulation and System Dynamics

6.1 Introduction

6.2 System Dynamics

6.3 Discrete-Event Simulation

6.4 Summary: The Basic Differences

6.5 Example: Modelling Emergency Care in Nottingham

6.6 The $64 000 Question: Which to Choose?

6.7 Conclusion

References

Chapter 7: Models as Interfaces

7.1 Introduction: Models at the Interfaces or Models as Interfaces

7.2 The Social Roles of Simulation

7.3 The Modelling Process

7.4 The Modelling Approach

7.5 Two Case Studies of Modelling Projects

7.6 Summary and Conclusions

References

Chapter 8: An Empirical Study Comparing Model Development in Discrete-Event Simulation and System Dynamics

8.1 Introduction

8.2 Existing Work Comparing DES and SD Modelling

8.3 The Study

8.4 Study Results

8.5 Observations from the DES and SD Expert Modellers' Behaviour

8.6 Conclusions

Acknowledgements

References

Chapter 9: Explaining Puzzling Dynamics: A Comparison of System Dynamics and Discrete-Event Simulation

9.1 Introduction

9.2 Existing Comparisons of SD and DES

9.3 Research Focus

9.4 Erratic Fisheries – Chance, Destiny and Limited Foresight

9.5 Structure and Behaviour in Fisheries: A Comparison of SD and DES Models

9.6 Summary of Findings

9.7 Limitations of the Study

9.8 SD or DES?

Acknowledgements

References

Chapter 10: DES View on Simulation Modelling: SIMUL8

10.1 Introduction

10.2 How Software Fits into the Project

10.3 Building a DES

10.4 Getting the Right Results from a DES

10.5 What Happens After the Results?

10.6 What Else does DES Software do and Why?

10.7 What Next for DES Software?

References

Chapter 11: Vensim and the Development of System Dynamics

11.1 Introduction

11.2 Coping with Complexity: The Need for System Dynamics

11.3 Complexity Arms Race

11.4 The Move to User-Led Innovation

11.5 Software Support

11.6 The Future for SD Software

References

Chapter 12: Multi-Method Modelling: AnyLogic

12.1 Architectures

12.2 Technical Aspect of Combining Modelling Methods

12.3 Example: Consumer Market and Supply Chain

12.4 Example: Epidemic and Clinic

12.5 Example: Product Portfolio and Investment Policy

12.6 Discussion

References

Chapter 13: Multiscale Modelling for Public Health Management: A Practical Guide

13.1 Introduction

13.2 Background

13.3 Multilevel System Theories and Methodologies

13.4 Multiscale Simulation Modelling and Management

13.5 Discussion

13.6 Conclusion

References

Chapter 14: Hybrid Modelling Case Studies

14.1 Introduction

14.2 A Multilevel Model of MRSA Endemicity and Its Control in Hospitals

14.3 Chlamydia Composite Model

14.4 A Hybrid Model for Social Care Services Operations

References

Chapter 15: The Ways Forward: A Personal View of System Dynamics and Discrete-Event Simulation

15.1 Genesis

15.2 Computer Simulation in Management Science

15.3 The Effect of Developments in Computing

15.4 The Importance of Process

15.5 My Own Comparison of the Simulation Approaches

15.6 Linking System Dynamics and Discrete-Event Simulation

15.7 The Importance of Intended Model Use

15.8 The Future?

References

Index

End User License Agreement

List of Tables

Table 2.1

Table 2.2

Table 2.3

Table 2.4

Table 2.5

Table 2.6

Table 4.1

Table 6.1

Table 6.2

Table 6.3

Table 6.4

Table 7.1

Table 7.2

Table 7.3

Table 8.1

Table 8.2

Table 8.3

Table 8.4

Table 8.5

Table 9.1

Table 9.2

Table 9.3

Table 13.1

Table 14.1

Table 14.2

List of Illustrations

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

Figure 2.6

Figure 2.7

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

Figure 3.8

Figure 3.9

Figure 3.10

Figure 3.11

Figure 3.12

Figure 3.13

Figure 3.14

Figure 3.15

Figure 3.16

Figure 4.1

Figure 4.2

Figure 4.3

Figure 5.1

Figure 6.1

Figure 6.2

Figure 6.3

Figure 6.4

Figure 6.5

Figure 7.1

Figure 8.1

Figure 8.2

Figure 8.3

Figure 8.4

Figure 8.5

Figure 9.1

Figure 9.2

Figure 9.3

Figure 9.4

Figure 9.5

Figure 9.6

Figure 9.7

Figure 9.8

Figure 9.9

Figure 9.10

Figure 9.11

Figure 9.12

Figure 9.13

Figure 9.14

Figure 9.15

Figure 10.1

Figure 10.2

Figure 10.3

Figure 10.4

Figure 10.5

Figure 10.6

Figure 10.7

Figure 10.8

Figure 10.9

Figure 10.10

Figure 10.11

Figure 10.12

Figure 10.13

Figure 10.14

Figure 11.1

Figure 11.2

Figure 11.3

Figure 11.4

Figure 11.5

Figure 11.6

Figure 11.7

Figure 11.8

Figure 11.9

Figure 11.10

Figure 11.11

Figure 11.12

Figure 11.13

Figure 11.14

Figure 11.15

Figure 11.16

Figure 11.17

Figure 11.18

Figure 11.19

Figure 11.20

Figure 11.21

Figure 11.22

Figure 11.23

Figure 11.24

Figure 11.25

Figure 11.26

Figure 11.27

Figure 11.28

Figure 12.1

Figure 12.2

Figure 12.3

Figure 12.4

Figure 12.5

Figure 12.6

Figure 12.7

Figure 12.8

Figure 12.9

Figure 12.10

Figure 12.11

Figure 12.12

Figure 12.13

Figure 12.14

Figure 12.15

Figure 12.16

Figure 12.17

Figure 12.18

Figure 12.19

Figure 12.20

Figure 12.21

Figure 12.22

Figure 12.23

Figure 12.24

Figure 12.25

Figure 13.1

Figure 13.2

Figure 13.3

Figure 13.4

Figure 14.1

Figure 14.2

Figure 14.3

Figure 14.4

Figure 14.5

Figure 14.6

Figure 14.7

Figure 14.8

Figure 14.9

Figure 14.10

Figure 15.1

Figure 15.2

Figure 15.3

Figure 15.4

Figure 15.5

Guide

Cover

Table of Contents

Preface

Chapter 1

Pages

ii

iii

iv

xv

xvi

xvii

xviii

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

Wiley Series inOperations Research and Management Science

Operations Research and Management Science (ORMS) is a broad, interdisciplinary branch of applied mathematics concerned with improving the quality of decisions and processes and is a major component of the global modern movement towards the use of advanced analytics in industry and scientific research. The Wiley Series in Operations Research and Management Science features a broad collection of books that meet the varied needs of researchers, practitioners, policy makers, and students who use or need to improve their use of analytics. Reflecting the wide range of current research within the ORMS community, the Series encompasses application, methodology, and theory and provides coverage of both classical and cutting edge ORMS concepts and developments. Written by recognized international experts in the field, this collection is appropriate for students as well as professionals from private and public sectors including industry, government, and nonprofit organization who are interested in ORMS at a technical level. The Series is comprised of three sections: Decision and Risk Analysis; Optimization Models; and Stochastic Models.

Advisory Editor • Decision and Risk Analysis

Gregory S. Parnell, United States Air Force Academy

Founding Series Editor

James J. Cochran, Louisiana Tech University

Discrete-Event Simulation and System Dynamics for Management Decision Making

Editors

Sally Brailsford

Southampton Business School, University of Southampton, UK

Leonid Churilov

Florey Institute of Neuroscience and Mental Health, Melbourne, Australia RMIT University, Melbourne, Victoria, Australia

Brian Dangerfield

Salford Business School, University of Salford, UK

This edition first published 2014

© 2014 John Wiley & Sons, Ltd

Registered office

John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom

For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com.

The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising herefrom. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Library of Congress Cataloging-in-Publication Data

Discrete-event simulation and system dynamics for management decision making / [edited by] Sally Brailsford, Leonid Churilov, Brian Dangerfield.

pages cm

Includes bibliographical references and index.

ISBN 978-1-118-34902-1 (hardback)

1. Discrete-time systems–Simulation methods. 2. System analysis. 3. Decision making. 4. Management science. I. Brailsford, Sally. II. Churilov, Leonid. III. Dangerfield, Brian Thornley.

T57.62.D495 2014

658.4′0352–dc23

2013047217

A catalogue record for this book is available from the British Library.

ISBN: 978-1-118-34902-1

Dedication

We dedicate this book to Ruth Davies, Emeritus Professor of Operational Research at the University of Warwick, a great academic and a great friend.

Preface

The genesis of operational research (OR) in the Second World War was largely characterised by deterministic techniques with a nod to risk evaluations such as in establishing the optimum balance of merchant ships and naval protection vessels in Atlantic convoy sizes. But it was part of the promulgation of OR techniques in large nationalised industries in the late 1940s and early 1950s that simulation came to the fore. This was particularly evident in the British steel industry. The emerging power of digital computers helped enormously and, under the guidance of luminaries such as Keith Tocher, discrete-event simulation (DES) (and the three-phase system) emerged from what had previously been Monte Carlo simulation.

Later in the 1950s in the United States another luminary, Jay Forrester, was settling into a new role at MIT and he saw the possibilities of applying the ideas and concepts from control engineering to the simulation of economic and social systems. Like Tocher, he relied on the growing power of computers. In fact he had been closely involved on the hardware side even to the extent of holding a US patent for random-access magnetic core memory. Forrester launched the field of what was to be system dynamics (SD), then known as industrial dynamics, in a paper in the Harvard Business Review in 1958.

Two powerful intellects were responsible for setting in train two separate methodologies in the domain of management science (MS) that, over the subsequent decades, have come to be employed on an enormous variety of applications. But although the simulation landscape has been enriched by their respective capabilities (which were rendered all the more impressive by the advent of icon-based computing) there has been almost no significant attempt to set out their respective merits in a comparative sense, still less to illustrate how they may be used in concert. This book, we believe, is the first volume to address these issues, while also describing agent-based (AB) modelling, another methodology that has recently emerged.

SCBLCBCD

List of Contributors

Steffen Bayer, Program in Health Services & Systems Research, Duke-NUS Graduate Medical School Singapore

Tim Bolt, Faculty of Health Sciences, University of Southampton, UK

Andrei Borshchev, Managing Director and CEO, The AnyLogic Company, St Petersburg, Russia

Sally Brailsford, Southampton Business School, University of Southampton, UK

Leonid Churilov, Florey Institute of Neuroscience and Mental Health, Melbourne, Australia; RMIT University, Melbourne, Victoria, Australia

Brian Dangerfield, Salford Business School, University of Salford, UK

Shivam M. Desai, Southampton Business School, University of Southampton, UK

Mark Elder, SIMUL8 Corporation, Glasgow, UK

Andrew Flitman, Florey Institute of Neuroscience and Mental Health, Melbourne, Victoria, Australia

Paul Harper, School of Mathematics, Cardiff University, UK

Lee Jones, Director, Ventana Systems UK, Oxton, Merseyside, UK

Maria Kapsali, Umea School of Business and Economics, Umea University, Sweden

Kathy Kotiadis, Kent Business School, University of Kent, Canterbury, UK

Geoff McDonnell, Centre for Health Informatics, Australian Institute of Health Innovation, University of New South Wales, Sydney, New South Wales, Australia; Adaptive Care Systems, Sydney, New South Wales, Australia

John Mingers, Kent Business School, University of Kent, Canterbury, UK

John Morecroft, London Business School, London, UK

Michael Pidd, Lancaster Business School, University of Lancaster, UK

Stewart Robinson, School of Business and Economics, Loughborough University, UK

Kristian Rotaru, Department of Accounting, Monash University, Melbourne, Victoria, Australia

Rosemarie Sadsad, Centre for Infectious Diseases and Microbiology – Public Health, Westmead Hospital, Sydney, New South Wales, Australia; Sydney Medical School, Westmead, The University of Sydney, New South Wales, Australia; Centre for Health Informatics, Australian Institute of Health Innovation, University of New South Wales, Sydney, New South Wales, Australia

Antuela Tako, School of Business and Economics, Loughborough University, UK

Joe Viana, Southampton Business School, University of Southampton, UK

1Introduction

Sally Brailsford,1 Leonid Churilov2 and Brian Dangerfield3

1Southampton Business School, University of Southampton, UK

2Florey Institute of Neuroscience and Mental Health, Melbourne, Australia; RMIT University, Melbourne, Victoria, Australia

3Salford Business School, University of Salford, UK

1.1 How this Book Came About

To begin at the end … the final chapter in this book, by Michael Pidd, contains both a backwards and a forwards look at system dynamics and discrete-event simulation. Historically, both modelling approaches originate from around the same time, the late 1950s and early 1960s. However, over the intervening decades they developed into separate scientific and practitioner communities, each with its own learned societies, academic journals and conferences. Discrete-event simulation (DES) has been a core subject on MSc programmes in operational research (OR) or management science (MS) from the 1970s onwards, and is a standard technique in the ‘OR/MS toolkit’. For many operational researchers, ‘simulation’ is synonymous with DES, and indeed the aim of the UK OR Society's own Journal of Simulation (quoting directly from the journal's web site) is to provide ‘a single source of accessible research and practice in the fast developing field of discrete-event simulation’ (http://www.palgrave-journals.com/jos/index.html). However, this is not true of system dynamics (SD): the SD community was, and still is to some extent, distinct from the OR community. While there is obviously some overlap in membership, there are many members of the international SD Society who are not members of their national OR Society (and vice versa). SD was certainly not taught on the MSc in OR at the University of Southampton in the 1980s and 1990s.

In 2000, the Simulation Special Interest Group of the UK OR Society held a joint meeting with the UK Chapter of the SD Society, entitled ‘Never the Twain Shall Meet’. At this meeting David Lane presented a paper (Lane, 2000) in which he discussed the differences between SD and DES and posed the question about whether they were ‘chalk and cheese’ or were actually two sides of the same coin. This meeting led to the foundation of a new OR Society Special Interest Group called ‘SD+’ whose aim was to bring the SD and OR communities together. The ‘+’ in SD+ was broader than just DES: it included many other OR techniques and approaches with which SD could interface. The ‘Never the Twain’ meeting also led to a number of academic papers exploring the similarities and differences between DES and SD, including the well-known study by Robinson and Morecroft which forms Chapter 9 of this book. Indirectly, it led to this book itself!

In some application areas the use of SD has expanded rapidly since 2000, and healthcare – the specialist application field of all three editors of this book – is one such area. However, despite initiatives like SD+, it is still true to say even today that SD is less well known in the mainstream OR community than DES. The number of DES papers at the annual Winter Simulation Conference, the major US conference on simulation, always greatly exceeds the number of SD papers. The main aim of this book is to begin to address this disparity. The book provides an integrated overview of SD and DES, a detailed comparison of the two approaches from a variety of perspectives, and a practical guide to how both may be used, either separately or together.

1.2 The Editors

As editors, we should declare our own personal interests. Having started out in the early 1990s as a dedicated user of DES, Sally Brailsford became interested in SD as a result of the ‘Never the Twain’ meeting, and joined David Lane in co-founding the SD+ group. Subsequently she became a zealous convert, using SD for several modelling studies. Like many other researchers, she was fascinated by the relationship between DES and SD, and in particular in the domain of healthcare (Brailsford and Hilton, 2001), but first used SD in practice in a project to model demand for emergency healthcare in Nottingham, England (Brailsford et al., 2004) which is described in detail in Chapter 6.

Leonid Churilov has a firm belief that real management problems do not come cleanly separated by disciplinary lines and, as a result, can rarely be comprehensively addressed using a single given modelling method. This basic premise is the source of continuous motivation for his keen interest in combining and contrasting different OR/MS techniques for management decision support. His research, in particular, included combining DES and clustering/classification techniques for decision support in hospital emergency departments, the use of both DES and SD for process systems modelling, and the original value-focused process engineering methodology that integrates the approaches from both the decision sciences and business process modelling domains. His work on philosophical underpinnings of both DES and SD worldviews from the critical realist perspective is featured in Chapter 5.

Brian Dangerfield discovered SD (industrial dynamics as it then was) over 40 years ago after a period in an OR unit in industry. Working at the (then) University of Liverpool School of Business Studies, he was a researcher on a project looking at the role of stocks in the UK economy. Rather than take the obvious econometric route he realised that an understanding of the macro role that stocks played in the workings of the economy would be much enhanced if, instead, macro-economic models were SD simulation models. Variables representing stocks had to be divorced from the flows which changed them. He was impressed with the way (i) SD models were forged at the policy level, (ii) separated out resource flows from the information flows driving changes in those resources and (iii) were also able to embrace relevant ‘soft’ variables which, using another methodology, might be excluded altogether. In sum he concluded that, given his overall knowledge and real-world experience with OR (including traditional simulation techniques), the SD methodology was just about the most promising in the landscape of OR, and his subsequent research career has concentrated on its use and development.

1.3 Navigating the Book

To our knowledge this book is unique – there is no similar coverage in one single volume. Books which cover both methodologies (e.g. Pidd's Computer Simulation in Management Science, 2009) merely split the page coverage – there is no attempt at integration, or a detailed comparison and description of how both approaches may be used in the same project. This book provides a seamless treatment of a variety of topics: theory, philosophy, detailed mechanics and practical implementation, all written by experts in the field. While some chapters are aimed at beginners, others are more advanced; the book also includes three software chapters which are very practical in nature.

The book is structured in seven unequal sections, three of which only contain one chapter. This Introduction forms the first section, and Pidd's concluding chapter the seventh. The second section, ‘Primers’, contains two chapters which provide a basic introduction to DES and SD respectively. These chapters are both written by experts with many years' experience of teaching and using each technique: Chapter 2 on DES (Stewart Robinson) and Chapter 3 on SD (Brian Dangerfield). The authors assume no prior knowledge of either technique, or an academic background in mathematics, statistics or OR. They are aimed at students and practitioners alike. The aim of both primers is to provide sufficient understanding to enable the average reader to get a basic grasp of the topic and be able to appreciate the subsequent chapters. The primers do not attempt to provide the breadth and depth of material offered in a typical MSc course in SD or DES, and they do not contain a great deal of technical detail. References are provided for anyone who does wish to delve deeper into the technicalities!

The third section also consists of a single chapter. By way of contrast with Chapter 2 and 3, the authors of Chapter 4 (Kathy Kotiadis and John Mingers) take a strongly academic stance. This chapter provides a theoretical background for much of the discussion in later chapters about combining different modelling paradigms. This chapter, which was originally published as a research article in the Journal of the Operational Research Society, discusses the combination of problem structuring methods with hard OR methodologies. Kotiadis and Mingers reflect on the barriers to such combinations that can be seen at the philosophical level – paradigm incommensurability – and cognitive level – type of personality and difficulty of switching paradigm. They then examine the combination of soft systems methodology and DES within a healthcare case study. They argue, by way of the practical application, that these problems are not insurmountable and that the result can be seen as the interplay of the soft and hard paradigms. The idea of yin and yang is proposed as a metaphor for this process.

The fourth section, ‘Comparisons’, is by far the largest and represents the heart of the book – its raison d'être. It contains five chapters, all of which consider contrasting aspects of DES and SD, ranging from methodological and philosophical comparisons using different lenses or frameworks, through to practical aspects and software implementations. In Chapter 5, Leonid Churilov, Kristian Rotaru and Andrew Flitman investigate how the critical realist philosophy of science facilitates explicit articulation of the fundamental philosophical assumptions underlying the SD and DES worldviews, using a practical illustration of simulating process systems. The ultimate aim is to achieve more effective use of simulation for intelligent thinking about, and management decision support in, process systems. The novelty and original contribution of this research is in applying the stratified ontology of critical realism, and the abductive mode of knowledge generation, to examine explicitly the philosophical bases of the DES and SD simulation worldviews. The outcomes of this research are targeted at both the manager, who is the contributor to, as well as the end user of, a simulation model of a real-world process system and, as such, could benefit from a clear understanding of how management knowledge is generated through the modelling process, and the management scientist who chooses to use simulation modelling to support management decision making in real-world process systems, and requires in-depth understanding of the scientific bases of the respective modelling methodologies to apply them in a truly scientific manner.

It is an oft-quoted cliché that if all you have is a hammer, then every problem is a nail. The aim of Chapter 6 is to discuss whether the choice of simulation methodology – DES or SD – is purely down to the personal preference and expertise of the modeller, or whether there are identifiable features of certain problems that make one approach intrinsically preferable to the other. Although from a methodological standpoint the overall comments are generic and applicable to any setting, the chapter has a bias towards healthcare applications as this is the area of domain expertise of the author, Sally Brailsford. A case study in emergency care is presented, where both DES and SD were used to tackle different aspects of the overall problem. The chapter concludes with some general guidelines to assist the modeller in making the choice of technique.

One commonly held stereotype is that SD naturally lends itself to problems where a group of people with potentially conflicting objectives need to be engaged, on an ongoing basis, in the process of developing and running a simulation model, whereas DES is more naturally suited to problems where there is an agreed objective and, after an initial meeting with the client, the modeller goes away, locks him- or herself in a darkened room and spends a month writing code, requesting data, running the model and obtaining results, which the modeller then presents to the client at the next meeting. This is clearly rather an absurd exaggeration, but like all stereotypes it is worth examining to see whether it contains a grain of truth. The use of DES and SD models in group model building projects, where a group of domain experts or other stakeholders come together to build a model together with a modelling expert, is examined in Chapter 7. Steffen Bayer, Timothy Bolt, Sally Brailsford and Maria Kapsali show how both SD and DES models have a social function and can act as an ‘interface’ between participants in a modelling project. An interface can be understood as a point of interaction which allows two systems to communicate across a boundary. In everyday use ‘interface’ is, however, also often used as relating to what affords and enables the process of communication across the boundary: interface as the means or mechanism that allows the boundary between subsystems to be permeable. The metaphor ‘interface’ highlights the potential of models (and of the modelling process) to support information transmission in the widest sense between the participants in a group modelling project but also across the boundary between the model and those engaging with the model. The aim of this chapter is to explore and illustrate how the metaphor of interface can shed light on the variety of social functions models can have, especially in a group modelling context.

The actual model building process is explored in further detail in the following chapter, which focuses less on the social aspects of model building and more on the technical aspects, providing a detailed comparison of DES and SD by analysing the thought processes of the modellers themselves. In Chapter 8 Antuela Tako and Stewart Robinson describe an empirical study which used verbal protocol analysis to study the modelling process followed by ten expert modellers, five using SD and five using DES. The participants were asked to build simulation models based on a case study and to think aloud while modelling. The generated verbal protocols are divided into seven modelling topics (problem structuring, conceptual modelling, data inputs, model coding, validation and verification, results and experimentation, and implementation) and are then analysed. Quantitative analysis of the verbal protocols shows that all modellers switch between modelling topics; however, DES modellers follow a more linear progression. They focus significantly more on model coding and verification and validation, whereas SD modellers focus on conceptual modelling. Observations are also made, revealing some interesting differences in the way the two groups of modellers tackle the case. This chapter contributes to the comparison of DES and SD modelling in management decision making by providing empirical evidence with regards to the model development process.

The ‘Comparisons’ section concludes with Chapter 9, a contribution from John Morecroft and Stewart Robinson. Morecroft (an expert in SD) and Robinson (an expert in DES) compare SD and DES through an examination of the way each method analyses the puzzling dynamics inherent in a fisheries system. They begin by recounting comparisons between SD and DES made by other authors over the past 20 years or so. They lament that none of these comparisons are offered by neutral parties – all have a main affiliation with either DES or SD.

Morecroft and Robinson start their own comparison by presenting data charts from two well-known fisheries. In one a serious drop in annual fish ‘landings’ recovered eventually, but in the other the new rate of substantially lower catches has continued for over 40 years. A series of experiments are then undertaken involving building dual models of a stylised fishery – one in SD and the other in DES. The authors compare, step by step, the two emerging models, the equation formulations and the resultant simulated behaviour. They begin with a ‘natural’ fishery where there is no human involvement and then progress to a harvested fishery where ships are involved and the fish are landed. Initially the harvested fishery is modelled in equilibrium but this assumption is ultimately relaxed – in the SD model by adding a nonlinearity and in the DES model by adding randomness. The SD model exhibits growth initially but ultimately the fishery collapses. Turning to the DES model, there is still a collapse but over a different timescale: the greater the spread of the normal distribution generating the randomness, the faster the eventual collapse. The authors conclude with a final experiment where the harvested fishery now includes endogenous investment in ships. The authors demonstrate that, in both models, the policies for deciding on ship investments are defective. In conclusion they express the paradigmatic differences between the two approaches by stating that DES illuminates interconnected randomness whereas SD illuminates deterministic complexity.

Underpinning the developments in both DES and SD modelling, the improvements in software functionality stand supreme. As Pidd describes in Chapter 15, probably until well into the 1980s both these methodologies were employed in a manner which today's user would find almost unbelievable. Text-based interfaces were the norm and the mainframe-supported software inevitably involved significant delays in delivering (usually printed) output. Common in DES at the time were such packages as GPSS and ECSL; in SD, DYNAMO and DYSMAP. All this changed by the mid-1990s as icon-based software became the norm, deployed on a PC and, latterly, on powerful laptop computers. Now users could experience almost instantaneous output together with the ability to display animated dynamic visual representations of their systems in real time.

This volume would be deficient if it did not afford some space to the consideration of contemporary software offerings. Accordingly, in the fifth section three authors have been invited to write about the software systems with which they are associated. Two were selected as being examples of commonly used software products for the methodologies covered by this volume and which had been in existence for at least 10 years to date; the third is a relative newcomer. The purpose of this book is such that an excessively lengthy documentation of relevant current software would be counter-productive. There are certainly other available systems for both DES and SD simulation and there is no intention to imply that those described here are in any sense superior; they are purely representative. And, going forward from here, Pidd ponders on possible future software developments in Chapter 15.

A word about the inclusion of AnyLogic is necessary as its emergence is relatively recent. Developments in the two principal methodologies covered here (and those related to them) are continuing apace. However, agent-based modelling (ABM) has emerged as a viable alternative to DES or SD simulation. Its use is primarily for certain types of problem, such as where a high level of (network) interaction detail and granularity is desirable and is capable of being represented as a feedback system. ABM would not exist if it were not for the astonishing developments in computer power in recent years. The AnyLogic system claims to allow for coding a problem as a DES, SD or ABM model. Whether one would wish to replicate a problem system in this way is open to debate. Usually the purpose of the investigation will dictate the approach adopted in coding the model. However, the availability of software such as AnyLogic means that, in theory, one would not need to utilise any other software system in order to create a methodologically diverse variety of systems models and, furthermore, it is possible to combine any of the three methods when using this software.

In Chapter 10 Mark Elder introduces a popular DES software platform – SIMUL8. He begins by emphasising that the model can aid understanding for the client such that he or she can have the confidence to change the process being managed for the better. Even rudimentary models can generate insight and a high-level model can produce the spark which spurs further investigation using more detailed models. While the thrust of this centres upon SIMUL8's dynamic ‘floor plan’ capabilities, and the way that this on-screen animation hooks in the client group, some of Elder's ideas about the benefits derived from the contemporary modelling process, rather than the model itself, would apply equally to SD.

He continues by taking the reader through an example model using the SIMUL8 software, depicting the processes involved in manufacturing metal tables. Some data can be automatically inserted by the software, thereby speeding up model creation. The need for replicated runs (trials) is emphasised together with the concomitant statistical analysis of the data generated. Some attention is given to the importance of verification and validation of the model as well as how to optimise the resources in the model. Finally, he concludes by considering the future for DES software.

Lee Jones continues the discussion of software matters by describing the Vensim SD simulation platform in Chapter 11. He starts by recounting how complexity in the modern business and social environment might be conducive to the growing use of SD. But this has not happened (particularly in business) and he conjectures that at its heart the issue may be bound up with SD software. Tracing the development of SD software platforms since the text-based environments of the 1980s, he recounts how Vensim's functionality has been improved over the years and how software user forums can now dictate the direction of additions to software functionality.

The importance of model units checks is underlined through the story of how NASA's Mars Climate Orbiter was destroyed on entry to the Martian atmosphere, merely because of a mismatch between metric and imperial units in the software driving the craft's orbital trajectory. Jones further considers calibration of a model to past (and indeed future) data using Vensim's optimisation feature. Other developments mentioned include the ‘Reality Check’ facility, together with ‘Story Telling’, the often problematic phase of SD modelling where the reasons for model behaviour have to be effectively communicated to the client; reduced form model diagrams, causes trees and strip graphs can all be usefully deployed to this end. Sensitivity graphs and policy optimisations can require many thousands of repeated runs and modern SD software harnesses the raw computing power now available in order to provide features of this sort. The chapter concludes by recounting how innovations in technology are helping the enhanced exposure of SD models: from the promulgation of models running live over the Internet, to the Facebook interactive game about how best to develop our planet (‘Game Change Rio’); and, for the immediate future, the possible utilisation of tablet computers and smart phones for model execution and presentation.

In the final chapter in the ‘Software’ section (Chapter 12) Andrei Borshchev provides an account of the relatively recent offering: AnyLogic. His underlying theme is to illustrate how this platform can be employed to mix three simulation modelling methodologies – DES, SD and ABM – within the same overall problem. By introducing the concept of the statechart, Borshchev instances seven theoretical occurrences of how agents, discrete elements and SD stocks and flows can interact. Going into much more of the model formulation details, he then describes three specific examples: a supply chain where the market is handled using SD concepts while the supply chain itself involves the discrete-event methodology; an epidemic model exploring a clinic's capacity to cope, where the combination here is ABM and DES; and finally a product life cycle and new product investment policy where the formulation involves a mix of ABM and SD. However, it would be wrong to imply that only combinations of two methodologies are possible: the AnyLogic platform can be employed to develop simulation models where all three co-exist.

The sixth section, which contains two chapters, focuses on models and applications and begins with a practical guide to using different levels or scales of modelling. Chapter 13, by Geoff McDonnell and Rosemarie Sadsad, is dedicated to various aspects of multiscale modelling in the context of public health and health services management. Simulation models of complex and multilevel organisational systems like healthcare are often abstracted at one level of interest. There are many challenges with developing simulation models of systems that span multiple organisational levels and physical scales. Several theoretical and conceptual frameworks for multilevel system analysis are analysed and an approach for developing multiscale and multimethod simulation models to aid management decisions is presented. Simulation models are used to illustrate how management actions, informed by patterns in stock levels, govern discrete events and entities, which, collectively, change the flow mechanism that controls stock levels. The proposed approach explicitly considers the role of context when designing and evaluating, in particular, public health actions. It extends current analytical and experimental methods and has the potential to encourage more collaborative and multidisciplinary effort towards effective public health management.

Chapter 14 focuses on applications of combined simulation approaches, and its objective is to demonstrate the use of hybrid modelling for management decision support. The chapter presents three separate case studies that are unified both by the common theme of using different modelling techniques in a hybrid manner and by the health- and social care context. The first case study, by Geoff McDonnell and Rosemarie Sadsad, combines the use of SD and ABM to better understand and control the spread of a specific drug-resistant pathogen in hospitals; the second case study, by Joe Viana, demonstrates how a hybrid SD and DES model can support clinical and management decision making in the context of sexual health services in Portsmouth, UK; while the third case study, by Shivam Desai, is dedicated to the investigation of a hybrid model that consists of an SD-inspired cell-based population model and a DES model, used to explore the performance of a contact centre for long-term care for people aged 65 and over.

The final chapter is Michael Pidd's personal view of the future – and of the past. Pidd is of course well known as the author of one of the best known and most widely used textbooks on simulation, first published in 1984 and now in its fifth edition (Pidd, 2009). In the 1980s, general OR textbooks tended to have about 10 chapters on mathematical programming, 10 chapters on other analytical techniques, and then a final chapter on the statistical aspects of Monte Carlo simulation, only to be used when all else fails. No wonder that Pidd's highly readable, practical and comprehensive book on simulation became a classic. Moreover, unusually for a simulation textbook, even the first edition contained a section on SD. It is hard to think of a more appropriate person to conclude this volume, which we hope will prove useful to academics and practitioners alike.

References

Brailsford, S.C. and Hilton, N.A. (2001) A comparison of discrete event simulation and system dynamics for modelling healthcare systems. Proceedings of ORAHS 2000, Glasgow, Scotland (ed. J. Riley), pp. 18–39.

Brailsford, S.C., Lattimer, V.A., Tarnaras, P. and Turnbull., J.A. (2004) Emergency and on-demand health care: modelling a large complex system.

Journal of the Operational Research Society

,

55

, 34–42.

Lane, D.C. (2000) You just don't understand me: modes of failure and success in the discourse between system dynamics and discrete event simulation. LSE OR Dept Working Paper LSEOR 00-34.

Pidd, M. (2009)

Computer Simulation in Management Science

, 5th edn, John Wiley & Sons, Ltd, Chichester.

2Discrete-Event Simulation: A Primer

Stewart Robinson

School of Business and Economics, Loughborough University, UK

2.1 Introduction

Discrete-event simulation (DES) grew largely out of a desire to model manufacturing systems. Based upon the foundation of Monte Carlo methods, DES models were developed to improve the design and operation of manufacturing plants. Among the earliest examples is the work of K.D. Tocher who developed the General Simulation Program in the late 1950s at the United Steels Companies in the United Kingdom; see Hollocks (2008) for an excellent summary of these early developments.

Over the years DES has been applied to a much broader set of applications including health, service industries, transportation, warehousing, supply chains, defence, computer systems and business process management. Much of this work has focused on improving the design and operation of the systems under investigation, but there have also been examples of DES for aiding strategic decision making.

DES is seen as a, if not the, mainstream simulation approach in the field of operational research (OR). Indeed, many OR specialists simply refer to it as ‘simulation’, seemingly ignoring the potential to simulate using other approaches, including system dynamics. DES does imply a very specific approach to simulation from both a technical and a philosophical perspective. In this chapter we will explore both of these perspectives. To set a context, we first present an example of a DES model. We then discuss how DES works (the technical perspective) followed by the worldview adopted by DES modellers (the philosophical perspective).

2.2 An Example of a Discrete-Event Simulation: Modelling a Hospital Theatres Process

Figure 2.1 shows an example of a DES model, in this case of a hospital theatres process (Burgess et al., 2011). This model can be accessed from www.simlean.org (accessed May 2012) as either a model file (for which the proprietary software is required) or as video files. The model simulates the flow of patients through an outpatients theatre over the period of a day. Patients arrive at a reception area and having registered they are prepared for their operation. Following the operation in one of the four theatres, they are moved to a recovery area. From here they are generally discharged, but in some cases the patient needs to be admitted to a ward.

Figure 2.1A simple DES model of a hospital theatres process (Source: Burgess et al., 2011; Robinson et al., 2012). Reproduced by permission of Elsevier.

Figure 2.1 shows the state of the model at around 11:35 in the morning, after an 8 a.m. start. At this time the theatres and preparation area are both fully utilised, resulting in a queue of patients waiting post-reception. There is also a queue for discharge since there is insufficient resource available for discharging patients at this point in the day.

While the model runs it records data on the performance of the system. These performance data typically summarise the experience of the individuals in the system, in this case patients, or the experience of the activities (e.g. preparation, theatres) and resources (e.g. receptionist, nurses, doctors). We are specifically interested in the amount of time the individuals spend waiting versus being worked on, which must be balanced with the utilisation of the activities and resources. By nature, high utilisation tends to lead to more waiting as the system becomes overloaded, while low waiting times tend to result from underutilised resources. The aim is to find an appropriate balance between these two measures of performance.

Figure 2.2 shows an example of some results from the hospital theatres process model. This splits the patient experience into value (receiving attention or treatment), waiting and blocked (an activity is complete, but the patient cannot move on because there is no space downstream, e.g. unable to leave the theatre because recovery is full). The pattern that emerges is, because patients arrive in batches, patients at the tail of a batch have to wait much longer to be seen.

Figure 2.2Results generated from the DES model of a hospital theatres process (Source: Burgess et al., 2011). Reproduced with permission from SimLean Publishing.

There is clearly a problem with waiting and blocking in the theatres process as it stands. The simulation can be used to investigate alternative process designs, different ways of scheduling and allocating resources, and changes to the management of the process.

2.3 The Technical Perspective: How DES Works

In simple terms, DES models queuing systems as they progress through time. In doing so it represents the world as entities that flow through a network of queues and activities. Where resources are shared between activities, these are also represented in a DES model. As such, the fundamental building blocks of a DES model are as follows:

Entities

: individual items that flow through the system, for example widgets in a manufacturing plant, people in a service system or hospital, packets of information in a computer network, vehicles in a transport system, orders in a supply chain.

Queues

: areas where entities wait to be worked on, for example buffers and stores, inventory, waiting areas, waiting lists, phone call queuing.

Activities

: perform work on entities, for example machines, travelling, moving, serving, printing.

Resources

: required to be present to operate activities, for example operators, equipment, doctors.

This structure is very flexible and it can be used to model a wide variety of systems. It is for this reason that DES has been widely used in OR for the types of application listed in the introduction to this chapter.

Figure 2.3 shows an example of a queuing system. Entities arrive at queue 1 and are then directed to either activity 1a or 1b; this could be based on the individual requirements of the entity or at random. Resource 1 is required to be present while activity 1 is taking place, hence it is shared between activities 1a and 1b. Following activity 1, all entities go to queue 2 to await activity 2, after which they leave the system. This system could represent many forms of two-stage queuing process, for instance a doctor's surgery with a reception area (with two distinct checking-in processes) followed by a consultation with the doctor.

Figure 2.3An example of a queuing system.

An important feature of entities in DES models is that they have attributes that describe specific features of an entity. These can describe features such as entity type, dimensions, weight, priority, order number and time in system. The values of these attributes can vary from entity to entity. As such, the attributes can be used to determine the logic of the model, for instance the time an individual entity will spend in an activity, its priority in a queue, or its route through the system. This ability to model the detail of individual entities is a particularly powerful feature of DES models.

Beyond these fundamental building blocks, there are two other key elements of DES models: the time-handling mechanism and methods for random sampling. We will now explore both of these in turn.

2.3.1 Time Handling in DES

A simple way of modelling the progression of time in simulation models is to use a constant time step. This approach works well for modelling systems that change continuously; a fixed time step is used to update the state of a system that is continuously changing. Of course, this only approximates continuous change, but digital computers are not able to truly model continuous time. If the time step is sufficiently small, the approximation enables the system to be modelled with sufficient accuracy. There is a trade-off, however, between run-time and accuracy, both of which increase with smaller time steps.

Such ‘continuous’ simulation is used widely in, for instance, the physical sciences and engineering for modelling continuous processes such as wave motion or the aerodynamics of a vehicle. It is also used in a business context to model processes that are subject to continuous change, for instance chemical plants or oil pipelines. Inventory systems, including supply chains, can be modelled using this approach if we are satisfied with updating the state of the system at fixed intervals, say daily. System dynamics is itself a form of continuous simulation.

Given that DES focuses on modelling individual entities that flow through a set of queues and activities, the exact timing of a change in the state of the system is difficult to predict. While an entity is being worked on in an activity, the state of the system remains unchanged. It is only when the activity is finished that the system state changes; the entity moves to the next stage of the process and the activity can start work on another entity. To model this accurately might require that time is modelled in very fine-grained units. To achieve this using a continuous simulation approach would require a very small time step. Because the system state only changes relatively rarely in comparison with the scale of the time step, the simulation would be extremely inefficient and slow since the majority of time steps would model points at which there is no change in system state.

To overcome this problem, DES focuses only on time points where there is a change in the state of the system. We define events as discrete points in time at which the system state changes. Typical events are an entity arrives, an activity starts and an activity ends. Since these events occur at irregular intervals, the time in a DES model does not move forward in a regular fashion, but jumps forward irregularly between the times at which events occur. It is this focus on events that leads to the name discrete-event simulation; that is, the name of the approach is describing the manner in which time is handled.

Table 2.1 shows an example of a DES of the queuing system in Figure 2.3. The table describes the sequence of events (in italics) and the consequences of those events. For instance, the first entity (Entity 1) arrives at the start of the simulation (time = 0). Because Activity 1a is idle, the entity can pass straight through Queue 1 into the activity which can then start work. For simplicity, Resource 1 is not included in this simulation. The first column of the table shows how time moves forward in an irregular fashion. Entities 1 to 4 arrive at times 0, 2, 7 and 23 respectively. We will show how to model these irregular arrival times in the next subsection. Meanwhile, Activity 1a takes 11 minutes, Activity 1b takes 25 minutes and Activity 2 takes 19 minutes. There is no reason, of course, that the times have to be integer numbers. We would normally model time using real numbers to the level of precision that the data, and computer technology, will allow.

Table 2.1 Example of a DES of the queuing system in Figure 2.3.

Time (min)

Event

0

Entity 1

arrives

, passes through Queue 1 to Activity 1a, Activity 1a

starts

2

Entity 2

arrives

, passes through Queue 1 to Activity 1b, Activity 1b

starts

7

Entity 3

arrives

and waits in Queue 1

11

Activity 1a

finishes

, Entity 1 passes through Queue 2 to Activity 2, Activity 2

starts

, Entity 3 moves to Activity 1a, Activity 1a

starts

22

Activity 1a

finishes

, Entity 3 waits in Queue 2

23

Entity 4

arrives

, passes through Queue 1 to Activity 1a, Activity 1a

starts

27

Activity 1b

finishes

, Entity 2 waits in Queue 2

30

Activity 2

finishes

, Entity 1 leaves the system, Entity 3 moves to Activity 2, Activity 2

starts

There is a range of ways that are used for managing the progression of time in DES. This requires determining the time at which events are due to occur, maintaining a record of those events and then, at the appropriate time, executing the events along with any consequent changes to the system state. We will not discuss the time-handling mechanisms for DES in detail here. For those that are interested, there is a very good description of these in Pidd (2004). For our purposes, it suffices to understand the key principle of moving time forward based on the instants at which discrete changes in system state (events) occur, which leads to an irregular time step.

2.3.2 Random Sampling in DES

A key element of almost all DES models is the need to represent randomness in the system under study, in recognition of the fact that many systems are by nature stochastic. Randomness occurs in the length of time an activity takes, for instance the time to serve a customer, the time to repair a piece of equipment and the time to process an order. It is also common for the arrival of entities to be subject to randomness, for example the arrival time of customers at a service system. Other factors are also random in nature, such as the size of an order, the weight of an item and the route an entity will choose to take through the system. In all these cases we might know how the data vary, in other words we know the distribution of the data, but we do not know what the exact value of the data will be in the next instance. So, for example, we might know that service time varies between 10 and 20 minutes, we might even know the exact shape of the distribution for service time, but we cannot predict what the exact service time for the next customer will be. To simulate systems that are subject to such randomness we need a mechanism for modelling this variability, for which Monte Carlo methods are employed. We now discuss how this is achieved in DES.

Say that we wish to simulate a ticketing web site and as part of that we want to model the number of tickets a customer will buy for a sports event. We may have looked at historic data for similar events and observed the distribution shown in Figure 2.4.

Figure 2.4