Explore the full ebook collection and download it now at ebookball.

com

Computer Simulation in Management Science 5th

Edition by Michael Pidd ISBN 0470092300
9780470092309

https://ebookball.com/product/computer-simulation-in-
management-science-5th-edition-by-michael-pidd-
isbn-0470092300-9780470092309-19802/

OR CLICK HERE

DOWLOAD EBOOK

Browse and Get More Ebook Downloads Instantly at https://ebookball.com

Click here to visit ebookball.com and download ebookball now
 Your digital treasures (PDF, ePub, MOBI) await
Download instantly and pick your perfect format...

Read anywhere, anytime, on any device!

Computer Simulation of Liquids 1st Edition by Michael

Allen, Dominic Tildesley ISBN 0198556454 9780198556459

https://ebookball.com/product/computer-simulation-of-liquids-1st-
edition-by-michael-allen-dominic-tildesley-
isbn-0198556454-9780198556459-13976/

ebookball.com

Lecture Notes in Computer Science 2145 1st edition by

Michael Leyton 3540427171 978‎-3540427179

https://ebookball.com/product/lecture-notes-in-computer-
science-2145-1st-edition-by-michael-
leyton-3540427171-978aeurz-3540427179-19932/

ebookball.com

Introduction To 80 86 Assembly Language And Computer

Architecture 1st Edition by Detmer ISBN 0763717738
9780763717735
https://ebookball.com/product/introduction-to-80-86-assembly-language-
and-computer-architecture-1st-edition-by-detmer-
isbn-0763717738-9780763717735-12404/

ebookball.com

Logic in Computer Science Modelling and Reasoning about

Systems 2nd Edition by Michael Huth, Mark Ryan ISBN
0511261586 9780511261589
https://ebookball.com/product/logic-in-computer-science-modelling-and-
reasoning-about-systems-2nd-edition-by-michael-huth-mark-ryan-
isbn-0511261586-9780511261589-19806/

ebookball.com
Lecture Notes in Computer Science 1600 1st edition by Ruth
Aylett, Michael Wooldridge, Manuela Veloso ISBN 3540664289
978-3540664284
https://ebookball.com/product/lecture-notes-in-computer-
science-1600-1st-edition-by-ruth-aylett-michael-wooldridge-manuela-
veloso-isbn-3540664289-978-3540664284-19660/

ebookball.com

Logic in Computer Science Modelling and Reasoning about

Systems 2nd Edition by Michael Huth, Mark Ryan ISBN
0511261586 9780511261589
https://ebookball.com/product/logic-in-computer-science-modelling-and-
reasoning-about-systems-2nd-edition-by-michael-huth-mark-ryan-
isbn-0511261586-9780511261589-19814/

ebookball.com

Introduction to 80 86 Assembly Language and Computer

Architecture 1st Edition by Richard C Detmer ISBN
0763746622 9780763746629
https://ebookball.com/product/introduction-to-80-86-assembly-language-
and-computer-architecture-1st-edition-by-richard-c-detmer-
isbn-0763746622-9780763746629-9016/

ebookball.com

Lecture Notes in Computer Science 1st Edition by Springer

ISBN

https://ebookball.com/product/lecture-notes-in-computer-science-1st-
edition-by-springer-isbn-10572/

ebookball.com

Modeling and Simulation in Science Engineering and

Technology 1st edition by Giovanna Guidoboni, Alon Harris,
Riccardo Sacco 3030258858 978-3030258856
https://ebookball.com/product/modeling-and-simulation-in-science-
engineering-and-technology-1st-edition-by-giovanna-guidoboni-alon-
harris-riccardo-sacco-3030258858-978-3030258856-20506/

ebookball.com
Computer Simulation in Management Science Fifth Edition

Computer Simulation in Management Science Fifth Edition Michael

Pidd Department of Management Science The Management School
Lancaster University John Wiley & Sons, Ltd

Copyright cg 2004 John Wiley & Sons Ltd, The Atrium, Southern
Gate, Chichester, West Sussex P019 8SQ, England Telephone (+44)
1243 779777 Email (for orders and customer service enquiries): cs-
[email protected] Visit our Home Page on www.wileyeurope.com or
www.wiley.com All Rights Reserved. No part ofthis publication may
be reproduced, stored in a retrieval system or transmitted in any
form or by any means, electronic, mechanical, photocopying,
recording, scanning or otherwise, except under the terms of the
Copyright, Designs and Patents Act 1988 or under the terms of a
licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham
Court Road, London WI T 4LP, UK, without the permission in writing
of the Publisher. Requests to the Publisher should be addressed to
the Permissions Department, John Wiley & Sons Ltd, The Atrium,
Southern Gate, Chichester, West Sussex P019 8SQ, England, or
emailed to [email protected], or faxed to (+44) 1243 770620.
This publication is designed to provide accurate and authoritative
information in regard to the subject matter covered. It is sold on the
understanding that the Publisher is not engaged in rendering
professional services. If professional advice or other expert
assistance is required, the services of a competent professional
should be sought. Other Wiley Editorial Offices John Wiley & Sons
Inc., III River Street, Hoboken, NJ 07030, USA Jossey-Bass, 989
Market Street, San Francisco, CA 94103-1741, USA Wiley-VCH Verlag
GmbH, Boschstr. 12, D-69469 Weinheim, Germany John Wiley &
Sons Australia Ltd, 33 Park Road, Milton, Queensland 4064, Australia
John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing
Distripark, Singapore 129809 John Wiley & Sons Canada Ltd, 22
Worcester Road, Etobicoke, Ontario, Canada M9WILI Wiley also
publishes its books in a variety of electronic formats. Some content
that appears in print may not be available in electronic books.
Library of Congress Cataloging-in-Publication Data Pidd, Michael.
Computer simulation in management science / Michael Pidd.-5th ed.
p.cm. Includes bibliographical references and index. ISBN 0-470-
09230-0 (pbk. : alk. paper) 1. Management science-Computer
simulatin. I. Title. T57.62.P532004 658.4'0352-dc22 2004003361
British Library Cataloguing in Publication Data A catalogue record for
this book is available from the British Library ISBN 0-470-09230-0
Project management by Originator, Gt Yarmouth, Norfolk (typeset in
1 O/12pt Photina) Printed and bound in Great Britain by TJ
International, Padstow, Cornwall This book is printed on acid-free
paper responsibly manufactured from sustainable forestry in which
at least two trees are planted for each one used for paper
production.

Contents Preface to the Fifth Edition xv PART I: FUNDAMENTALS OF

COMPUTER SIMULATION IN MANAGEMENT SCIENCE 1 1 The
computer simulation approach 3 1.1 Models, experiments and
computers 1.2 Some applications of computer simulation 1. 2 .1
Manufacturing 1.2.2 Health care 1.2.3 Business process re-
engineering 1.2.4 Transport systems 1.2.5 Defence 1.3 Models in
management science 1.4 Simulation as experimentation 1.5 Why
simulate? 1.5.1 Simulation versus direct experimentation 1.5.2
Simulation versus mathematical modelling 1.6 Summary Exercises
References 3 4 4 5 5 6 7 7 8 9 9 10 10 11 11 2 A variety of
modelling approaches 15 2.1 General considerations 15 2.2 Time
handling 15 2.2.1 Time slicing 15 2.2.2 Next-event technique 17
2.2.3 Time slicing or next event? 18 2.3 Stochastic or deterministic?
18 2.3.1 Deterministic simulation: a time-slicing example 19 2.3.2
Stochastic simulation 21

vi CONTENTS 2.4 Discrete or continuous change 2.4.1 Discrete

change 2.4.2 Continuous change 2.4.3 A few words on simulation
software Exercises References 25 26 26 27 27 28 3 Computer
simulation in practice 29 3.1 Process, content, problem and project
3.1.1 Process and content 3.1. 2 Problems and projects 3.1.3 Two
parallel streams 3.2 The simulation problem part of the study 3.3
Problem structuring 3.3.1 Problem structuring as exploration 3.4
Modelling 3.4.1 Conceptual model building 3.4.2 Computer
implementation 3.4.3 Validation 3.4.4 Experimentation 3.4.5
Implementation 3.5 The project part of the study 3.5.1 Initial
negotiation and project definition 3.5.2 Project management and
control 3.5.3 Project completion Exercises References 29 29 30 30
32 32 33 35 35 36 37 38 39 40 41 42 43 43 44 4 Static Monte Carlo
simulation 45 4.1 Basic ideas 4.1.1 Risk and uncertainty 4.1.2 The
replacement problem: a reprise 4.1.3 Static Monte Carlo simulation
defined 4.2 Some important considerations 4.2.1 Subjective
probabilities 4.2.2 Repeatability 4.3 Some simple static simulations
4.3.1 The loan repayment 4.3.2 An investment decision 4.4
Simulation on spreadsheets Exercises References 45 45 46 47 49 49
50 51 51 53 57 58 59

CONTENTS vii PART II: DISCRETE EVENT SIMULATION 61 5 Discrete

event modelling 63 5.1 Fundamentals 5.2 Terminology 5.2.1 Objects
of the system 5.2.2 The organization of entities 5.2.3 Operations of
the entities 5.3 Activity cycle diagrams 5.3.1 Example 1: a simple job
shop 5.3.2 Example 2: the harassed booking clerk 5.3.3 Example 3:
the delivery depot 5.3.4 Using the activity cycle diagram 5.4 Activity
cycle diagrams: a caveat Exercises References 63 63 64 65 65 66 68
71 74 77 78 79 81 6 How discrete simulation software worli.S 83 6.1
Introduction 83 6.1.1 Why understand how simulation software is
organized? 83 6.1.2 Simulation executives in more detail 84 6.1.3
Application logic 85 6.2 The three-phase approach 85 6.2.1 Bs 85
6.2.2 Cs 86 6.2.3 The exception to the general rule 87 6.2.4 Bs and
Cs in the harassed booking clerk problem 88 6.2.5 Another example:
aT-junction 89 6.3 How the three-phase approach works 90 6.3.1
The A phase 92 6.3.2 TheBphase 92 6.3.3 The C phase 92 6.4 The
harassed booking clerk-a manual three-phase simulation 93 6.4.1
The first A phase 94 6.4.2 The first B phase 94 6.4.3 The first C
phase 95 6.4.4 The second A phase 95 6.4.5 The next Band C
phases 96 6.4.6 The third A phase 96 6.4.7 The third B phase 96 6.5
The event-based worldview 97 6.5.1 Events in the harassed booking
clerk problem 97 6.5.2 Event-based executives 99

viii CONTENTS 6.6 The activity-scanning approach 100 6.6.1

Activities 100 6.6.2 Activity-scanning executives 101 6.7 Process-
based approaches 101 6.7.1 Processes in the harassed booking clerk
pro blem 102 6.7.2 Process interaction 103 6.7.3 Process-based
executives 104 6.8 Which approach is best? 105 6.8.1 Three-phase
versus process-based approaches 105 Exercises 106 References 108
7 Writing a three-phase simulation program 109 7.1 Introduction 109
7.1.1 The basic structure of the library 110 7.2 Inside the executive
III 7.2.1 The control array 112 7.2.2 Using the control array to
operate a three-phase simulation 112 7.3 The Visual Basic
implementation 113 7.3.1 Some comments on Visual Basic 113 7.3.2
The variables and their types 114 7.3.3 The A phase 116 7.3.4 The B
phase 118 7.3.5 The C phase 119 7.3.6 Running the simulation 119
7.4 Using VBSim to simulate the harassed booking clerk problem 120
7.4.1 Entities, Bs andCs 120 7.4.2 Personal enquirers and phone calls
arrive 121 7.4.3 The end of personal service and phone calls 123
7.4.4 Observations 124 7.4.5 The Cs 124 7.4.6 Initialization and
finalization 125 7.5 Putting it all together 126 Exercises 12 7
References 128 8 Visual interactive modelling and simulation 129 8.1
Basic ideas 8.1.1 Visual interactive modelling (VIM) 8.1.2 Visual
simulation output 8.1.3 Interaction 8.1.4 A caveat 8.2 Designing a
visual simulation display 8.2.1 Iconic displays 129 129 130 131 132
132 133

CONTENTS ix 8.2.2 Logical displays 134 8.2.3 Chart displays 135 8.3
VIMS 136 8.3.1 Joe's exhaust parlour 136 8.3.2 Joe's exhaust
parlour in Micro Saint: model building 138 8.3.3 Joe's exhaust
parlour in Micro Saint: running and analysing the simulation 140
8.3.4 Joe's exhaust parlour in SIMUL8: model building 143 8.3.5
Joe's exhaust parlour in SIMUL8: running and analysing the
simulation 145 8.4 Visual interactive simulation: a reprise 146
Exercises 146 References 146 9 Discrete simulation software 149 9.1
General principals 149 9.2 A quick overview of discrete simulation
software 150 9.3 VIMS and their relatives 151 9.3.1 VIMS-a reprise
151 9.3.2 Block diagram systems 152 9.3.3 VIMS and block diagram
systems 157 9.4 Programming using a general purpose language
158 9.4.1 Pros and cons 158 9.4.2 Libraries and component-based
software 159 9.5 Programming approaches using simulation
languages 161 9.5.1 Common features of simulation languages 162
9.5.2 An example: SIMSCRIPT 11.5 163 9.6 Layered systems and
application templates 168 9.6.1 Layered systems 168 9.6.2
Application templates 169 9.7 Appraising simulation software: some
principles 169 9.7.1 The type of application 170 9.7.2 The
expectations for end use 170 9.7.3 Knowledge, computing policy and
user support 171 9. 7.4 Price 1 71 9.8 Which to choose? Horses for
courses 172 9.8.1 VIMS 172 9.8.2 Simulation languages 172
References 1 73 10 Sampling methods 1 75 10.1 Basic ideas 175
10.1.1 General principles of random sampling 176 10.1.2 Top-hat
sampling 177

x CONTENTS 10.1.3 The fundamental random sampling process 178

10.1.4 Use of pre-written libraries of algorithms 179 10.2 Random
number generation 179 10.2.1 Truly random numbers 179 10.2.2
Pseudo-random numbers 180 10.2.3 Congruential generators 181
10.2.4 General requirements for these generators 182 10.2.5
Multiplicative congruential generators 182 10.2.6 Improving on
simple congruential generators 184 10.2.7 Using inbuilt random
number generators 185 10.3 Testing random number generators 185
10.3.1 Scatter plots 186 10.3.2 Auxiliarysequences 187 10.3.3
Frequency tests 187 10.3.4 Serial test 188 10.3.5 Gap test 188
10.3.6 Other tests 188 10.4 General methods for random sampling
from continuous distributions 189 10.4.1 Inversion 189 10.4.2
Rejection 192 10.4.3 Composition 193 10.5 Random sampling
algorithms for discrete distributions 193 10.5.1 Sampling from
histograms 193 10.5.2 Implicit inverse transformation 194 10.5.3
Discrete rejection-samples from a Poisson distribution 196 10.6
Sampling from the normal distribution 197 10.6.1 TheoriginalBox-
Miillermethod 198 10.6.2 Box-Miiller polar variation 199 10.6.3
Sampling from a normal distribution by composition 200 10.6.4 A
poor way to sample from the normal distribution 201 10.7 Deriving
one distribution from another-lag-normal variates 201 10.8 Sampling
from non-stationary processes: thinning 202 Exercises 203
References 204 11 Planning and analysing discrete simulation output
207 11.1 Fundamental ideas 11.1.1 Simulation as directed
experimentation 11.1.2 Estimation and comparison 11.1.3 Three
important principles 11.1.4 Some preliminary advice 11.2 Dealing
with transient effects 11.2.1 Terminating and non-terminating
systems 11.2.2 Achieving steady state 11.2.3 Using a run-in period
207 207 208 209 210 210 210 212 213

CONTENTS xi 11.2.4 Welch's method for determining the run-in

period 214 11.3 Dealing with lack of independence 215 11.3.1
Simple replication 215 11. 3 .2 Using batch means 216 11. 3.3
Overlapping batch means (OBM) 217 11.3.4 Regenerative methods
218 11. 4 Variance reduction 2 18 11.4.1 The basic problem-
sampling variation 219 11.4.2 Set and sequence effects 220 11.4.3
Common random number streams and synchronization 222 11.4.4
Control variates (regression sampling) 223 11.4.5 Antitheticvariates
225 11.5 Descriptive sampling 226 11.5.1 Basic idea 226 11.5.2
Procedure 227 11.6 Experimentation 228 11.6.1 Basic ideas 228
11.6.2 Factorial experiments 229 Exercises 23 1 References 23 1 12
Model Testing and Validation 233 12.1 The importance of validation
233 12.1.1 Validation is impossible, but desirable 233 12.1.2 Some
practical issues 234 12.1.3 The "real" world, the model and
observation 235 12.1.4 The hypothetic a-deductive approach 236
12.1.5 The importance of process and other aspects 237 12.2
Validation and comparison 237 12.2.1 Experimentalframes 238
12.2.2 Program verification and model validation 239 12.3 Black box
validation 240 12.3.1 Black box validation: a model's predictive
power 240 12.3.2 How valid? 240 12.3.3 Validation errors 241 12.3.4
Testing model components 241 12.4 White box validation 242 12.4.1
Detailed internal structure 242 12.4.2 Input distributions 242 12.4.3
Static logic 243 12.4.4 Dynamic logic 244 12.5 Type zero errors 245
12.5.1 Over-elaboration 245 12.5.2 Over-simplification 246 12.5.3
Steering a sensible course 246 References 246
xii CONTENTS PART III: SYSTEM DYNAMICS 247 13 Structure,
behaviour, events and Feedbacl{ systems 249 13.1 Events,
behaviours and structures 249 13.1.1 System simulation 249 13.1.2
The importance of system structure 2 50 13.2 Feedback systems 251
13.2.1 Hierarchical feedback systems: an example 251 13.2.2 Causal
loop diagrams 253 13.3 Modelling feedback systems 255 13.3.1
Delays 255 13.3.2 Levels and stocks 256 13.3.3 Rates and flows 257
13.3.4 Policies 258 13.4 The origings of system dynamics 258 13.4.1
Control theory 259 Exercises 260 References 261 14 System
dynamics modelling and simulation 263 14.1 Introduction 263 14.1.1
Stock and flow diagrams 264 14.1.2 A stock and flow diagram for
Big AI's problem 265 14.2 Beyond the diagrams-system dynamics
simulation 266 14.2.1 Time handling in system dynamics 267 14.2.2
Equation types 268 14.2.3 Powersim equations for Big AI's problem
269 14.2.4 Integration and the value of dt 270 14.3 Simulating
delays in system dynamics 272 14.3.1 Pipeline delays 272 14.3.2
Exponentialdelays 272 14.3.3 Information delays 274 14.4 System
dynamics modelling 274 14.4.1 Modelling from the outside in 275
14.4.2 Modelling from the inside out 276 Exercises 2 77 References
2 77 15 System dynamics in practice 279 15.1 Associated Spares Ltd
15.1.1 The problem as originally posed 15.1.2 The multi-echelon
system 15.1.3 The retail branch model 15.1.4 The regional
warehouse model 279 279 280 280 283

CONTENTS xiii 15.1. 5 The central warehouse model 284 1 5.1.6 The
total system model 28 5 15.1.7 Some conclusions 286 15.1. 8 A
postscript 287 15.2 Dynastat Ltd 288 15.2.1 An expansion
programme 288 15.2.2 The manpower problem 288 15.2.3
Recruitment 288 15.2.4 Turnover 289 15.2.5 Some effects of this
structure 290 15.2. 6 Validating the model 290 15.2.7 Simulation
results 291 15.2.8 Predicting length of service 292 15.2.9 The value
of the exercise to Dynastat 292 15.3 System dynamics in practice
293 15.3.1 Simple models 293 15.3.2 Communication 294 15.3.3
New thinking 294 15.3.4 Evolutionary involvement 295 References
295 Index 297
For SallYI Karen and Helen who have been a constant
encouragement during the two decades since I first proposed this
book.

Preface to the Fifth Edition As in the previous editions, this book is

aimed at management science students and practitioners who need
to learn how to conduct computer simulation studies. As before, its
main focus is on simulation modelling and all the material is
organized to that end. Since 1997, when I wrote the fourth edition,
progress in computing and in computer simulation has continued
apace. Hence the changes in this edition aim to keep the book up to
date, whilst making whatever other improvements seem sensible.
When I started work on the first edition in 1982 I had no idea that I
would be working on a fifth edition over 20 years later. Much as
software and hardware are upgraded every so often, so, it seems,
are books. A number of people have helped me in producing this
fifth edition by sug- gesting improvements to the third edition. They
include David Lane, Joyce Brown, Sarah Cope, Robert Fildes and
Nuno Melao. John Crookes, though now retired, stimulated many of
my original ideas and I am grateful to him for that. I acknowledge
the help and encouragement of Sarah Booth and her colleagues at
John Wiley & Sons and also the cooperation of Lanner Systems,
Micro Analysis and Design and the Simul8 Corporation. To all these,
thank you for your help-but I take the blame for any mistakes that
remain. Thanks to small laptop computers and the appalling
timekeeping of the Virgin West Coast train company, much more of
this edition has been written on trains than I had ever thought
possible. Perhaps if they made more use of the modelling
approaches described here, things would be better? As before, the
book is organized around three parts. PART I: FUNDAMENTALS OF
COMPUTER SIMULATION IN MANAGEMENT SCIENCE There are four
chapters in Part I and they provide a general introduction to the
principles of computer simulation. Chapter 3, Computer Simulation in
Practice suggests how a simulation study might be conducted.
Chapter 4 is wholly new and discusses the use of static Monte Carlo
methods in managing risk and uncertainty. The whole of Part I is
deliberately non-technical and makes little or no demand on
computing or statistical knowledge, other than the

xvi PREFACE TO THE FIFTH EDITION ability to use spreadsheets. It

serves as an introduction to those who wish to follow the rest of the
book in detail, but is also aimed at MBA and undergradu- ate
business majors who wish to gain an overview of the subject. PART
II: DISCRETE EVENT SIMULATION This is aimed at those readers
who need to know how to produce valid, working discrete event
simulation models. It covers four important aspects of discrete event
simulation methods. (1) Discrete event modelling: Chapter 5
introduces the general terminology of discrete event simulation and
shows, in some detail, how different ap- proaches may be
implemented. This is then continued in Chapter 6 to show how the
internals of discrete simulations can be programmed. Chapter 12
discusses the important issue of model testing and validation,
something that is so often squeezed out in practice. (2) Computing
aspects: Chapters 7, 8 and 9 cover different aspects of computing
that are related to discrete simulation. Chapter 7 shows how a three-
phase simulation model may be easily implemented in almost any
programming language. To support Chapter 7 I have made a set of
three-phase libraries available in C, C++, Visual Basic, Turbo Pascal
and Java. These can be found at http://www.lancs.ac.uk/staff/
smamp/. I retain copyright but accept no responsibility for their use
or misuse. Chapter 7 introduces two commonly used Visual
Interactive Modelling Systems-Micro Saint and SIMUL8. Chapter 8
provides a review of the main types of software available for discrete
event simulation. (3) Statistical aspects: Chapters 10 and 11 are
concerned with the statistical aspects of discrete event simulation
and form the final section of Part II. They describe how sampling
methods can be built into simulation models and how to disentangle
the problems of experimentation that follow in their wake. To follow
these chapters properly, the reader needs to understand basic
probability and statistics. (4) Chapter 12 discusses the important
issue of model testing and validation, something that is so often
squeezed out in practice. This chapter could also sit in Part III,
which is why it is placed at the end of Part II. PART III: SYSTEM
DYNAMICS The methods of system dynamics as first propounded by
Jay Forrester are, in my opinion, still the most widely used formal
simulation methods in manage- ment science after discrete event
methods. Hence I have devoted three chapters to the topic and have
attempted to provide a general introduction to its methodology in
Chapter 14. Chapter 15 discusses the detail of the approach and
Chapter 16 describes work carried out by Brian Parker in using the
methods to tackle managerial problems.

Part I

Fundamentals of Computer Simulation in Management Science

The Computer Simulation Approach

1.1 MODELS, EXPERIMENTS AND COMPUTERS

Management scientists are not easily separated from their computers

and for good reason. Since the 1960s, computers have become
smaller, cheaper, more powerful and easier to use by non-specialists.
In particular, the develop- ment of powerful and cheap portable
machines with excellent graphics has opened up wide areas of work
for the management scientist. Modern computers allow the analyst
to explore the whole range of feasible options in a decision problem.
These options could be explored without a computer but the process
would be very slow and the problem may well change significantly
before a satisfactory solution is produced. With a computer, large
amounts of data can be quickly processed and presented as a
report. This is extremely valuable to the management scientist. One
way in which a management scientist uses a computer is to simulate
some system or other. This is generally done when it is impossible or
inconvenient to find some other way of tackling the problem. In such
simulations, a computer is used because of its speed in mimicking a
system over a period of time. Again, most of these simulations could
(in theory at least) be performed without a computer. But in most
organizations, important problems have to be solved quickly: hence
the use of computer simulation in management science. Computer
simulation methods have developed since the early 1960s and may
well be the most commonly used of all the analytical tools of
management science. The basic principles are simple enough. The
analyst builds a model of the system of interest, writes computer
programs that embody the model and uses a computer to imitate
the system's behaviour when subject to a variety of operating
policies. Thus, the most desirable policy may be selected. For
example, a biscuit company may wish to increase the throughput at
a distribution depot. Suppose that the biscuits arrive at the depot on
large articulated trucks and, are unloaded and transferred onto
storage racks by forklift trucks. When required, the biscuits are
removed from the racks and loaded onto small delivery vans for
despatch to particular retail customers. To increase the throughput,
a number of options might present themselves to the management.
For example, they could:

4 THE COMPUTER SIMULATION APPROACH

(1) Increase the number ofloading or unloading bays. (2) Increase
the number of forklift trucks. (3) Use new systems for handling the
goods, etc.

It would be possible to experiment on the real depot by varying

some of these factors but such trials would be expensive and time
consuming. The simulation approach to this problem involves the
development of a model of the depot. The model is simply an
unambiguous statement of the way in which the various components
of the system (e.g., trucks and lorries) interact to produce the
behaviour of the system. Once the model has been translated into a
computer program the high speed of the computer allows a
simulation of, say, six months in a few moments. The simulation
could also be repeated with the various factors at different levels to
see the effect of more loading bays, for example. In this way, the
programmed model is used as the basis for experimentation. By
doing so, many more options can be examined than would be
possible in the real depot-and any disruption is avoided: hence the
attraction of computer simulation methods. To summarize, in a
computer simulation we use the power of a computer to carry out
experiments on a model of the system of interest. In most cases,
such simulations could be done by hand-but few people would wish
to do so. Now that computers offer significant power for a minimal
cost, a computer simulation approach seems to make even more
sense in management science.

1.2 SOME APPLICATIONS OF COMPUTER SIMULATION

Though it is impossible to be sure which techniques are most

commonly used in management science, the occasional surveys of
practitioners usually report simulation methods in the top three. This
section briefly reviews some of the main application areas.

1.2.1 Manufacturing

As markets for manufactured goods have become globalized,

manufacturers have increasingly attempted to mass customize their
products. That is, they have sought economies of scale by
developing products that will have global appeal and should sell in
many countries. At the same time they have had to ensure that the
products themselves are suited to local preferences, which means
they have had to produce local variants of the global designs. This
mass customization, sometimes known as glocalization, has placed
great pressure on manufacturers to develop and install
manufacturing systems that can deliver high volumes of high-quality
goods at low cost to meet local needs. This has led to huge
investments in manufacturing plant and asso- ciated control
systems. It is important to ensure that such systems operate as
intended, and therefore computer simulation methods have found an
important place in the process of designing and implementing these
manufac- turing plant and systems.

SOME APPLICATIONS OF COMPUTER SIMULATION 5

Examples of this use of computer simulation occur across most

manufacturing sectors and include food manufacturing (Pidd, 1987),
semi- conductor wafer fabrication (De Jong, 2001), beverages
(Harrell, 1993), automobile manufacture (Ladbrook and Januszczak,
2001), aerospace (Lu and Sundaram, 2002), shipbuilding (Williams
et aI., 2001) and materials handling (Burnett and LeBaron, 2001).
Simulation allows the comparison of alternative designs and control
policies on the model before starting to build the physical plant. It
helps to reduce the cost and risk of large-scale errors. Simulation
approaches are also used on existing plant to find better ways to
operate, and these studies might be one-off exercises or may be
part of a periodic check on the running of the system.

1.2.2 Health care

As with manufacturing, there is also a need to make effective use of

limited resources when providing and delivering health care. Thus,
simulation approaches have found widespread application in health
care systems around the world. Hupert et aI. (2002) discuss the
distribution of antibiotics and vaccines to dispensing centres in the
event of a terrorist attack. Ceric (1990) describes how the methods
were used to plan a system to move goods and equipment around a
large new hospital in an effective and efficient manner. McGuire
(1994) reports on the use of simulation for the planning of effective
emergency departments. In all such simulations, the idea was to test
different policies without putting patients to inconvenience or placing
them at risk. Simulation is also used to assess the effect of different
treatment pro- grammes. For example, Davies et aI. (2002) used
simulation to investigate ways in which the eyesight of diabetic
patients can be more effectively preserved and Jacobsen et aI.
(2001) used it to assess the value of paediatric immunization
programmes.

1.2.3 Business process re-engineering

Recent years have seen an increasing concern by businesses to
ensure that their core processes are operated effectively and
efficiently, and this has been the aim of business process re-
engineering (BPR). In BPR the idea is to take a fundamental look at
the basic processes without which the business could not function
and which contribute in a major way to both profit and cost. In some
ways, the stress on BPR mirrors the shift in manufacturing from
batch production towards flow-line manufacturing. An example of a
BPR exercise might be an investigation of the various operations and
activities involved in delivering goods to a customer, and in invoicing
that customer and in receiving payment. In a traditional system, the
paperwork and computer- based documentation might need to pass
through several different departments. Taking a radical look at such
processes might lead to great simplification and thus to reduced
costs and to better service. The aim of BPR is to take an integrated
view of such activities and to find ways to provide a better service at
lower cost by more effective organization.

6 THE COMPUTER SIMULATION APPROACH

Dennis et ai. (2000) summarize some applications in the telecomms

industry. Bhaskar et al. (1994) identify computer simulation as one
of the key approaches to understanding how business processes
might be re-engineered to improve performance. Davies (1994)
describes how simulation has been used in BPR in the UK financial
services industry. Companies providing these financial services in the
UK must meet legal time limits for their responses to customers and
must also carry out a series of checks required by law-in addition to
their own internal monitoring. Davies (1994) developed a simula-
tion model known as SCOPE to enable organizations to organze their
office processes so as to achieve target performance levels. SCOPE
works by simulat- ing the flow of documents through the
organization.

1.2.4 Transport systems

Computer simulation is also used in a wide range oftransportation

systems. As with other applications, the idea is to ensure that the
system operates as effi- ciently and as effectively as possible. In the
aviation sector, simulation methods have, for example, been used to
help plan large passenger terminals. Airport terminals include
systems for moving baggage and for ensuring that passengers can
get to the departure gates in time for their planes, and a number of
simulations have been used to assess their performance (e.g.,
Joustra and Van Dijk, 2001). Also in the aviation sector, air traffic
control systems are used to ensure that air space is used efficiently
and safely. As part of this, the air traffic controllers must ensure that
the movement of aircraft is planned in advance and then managed in
real time. Simulation approaches have made a great contribution to
safer and more cost-effective air traffic control (e.g., Lee et aI.,
2001). The shipping sector has also been a long-term user of
computer methods. Indeed, one of the computer simulation
programming languages (CSL; Buxton and Laski, 1962) was first
developed by Esso (predecessor of Exxon) to support simulations of
the movement of crude and refined oil around the world. Shipping
applications continue to this day and an example is given in Heath
(1993). Salt (1991) reports how simulation methods were used to
help plan the movement of traffic in the Channel Tunnel that links
the UK and France. Though the introduction of this service was
controversial, it is clear that one key to its future success is the
reliability ofthe service that it offers. Unlike the ferries which also ply
the route, bad weather should not prevent the operation of the
tunnel service. It was therefore crucial that its operations were
properly planned and managed. Salt (op cit) gives interesting
examples of how this was supported by computer simulation
approaches. The road transport sector is also a major user of
computer simulation methods both to plan individual companies'
operations and to investigate road traffic systems in general. Traffic
simulators are now a standard part of the armoury of road traffic
planners (Pidd, 1995; Rathi and Santiago, 1990; Young et aI., 1989)
since they permit possible road configurations and traffic
management schemes to be refined before their physical
implementation.

MODELS IN MANAGEMENT SCIENCE 7

1.2.5 Defence

The defence sector is also a major user of simulation methods and

the Proceed- ings of the Winter Simulation Conference usually
include a range of papers discussing work in this area (e.g., see
Robinson, 2001). Applications range from studies oflogistics
operations through to battle simulations, which inves- tigate possible
strategies and tactics to be used in defence or attack. Their appeal
here is obvious; no battle commander wishes to be defeated and the
chance to develop tactics beforehand and to prepare
countermeasures is of some importance. Not surprisingly, the
majority of defence simulations are not reported in the open
literature.

1.3 MODELS IN MANAGEMENT SCIENCE

Models of various types are often used in management science.
They are repre- sentations of the system of interest and are used to
investigate possible improvements in the real system or to discover
the effect of different policies on that system. This is not the place
for a detailed exposition of modelling; for this the reader should
consult Miser and Quade (1988), Pidd (2003), Rivett (1994), or
White (1975). However, some mention of the topic is necessary. The
simplest type of model employed in management is probably a scale
model, possibly of a building. By using scale models it is possible to
plan sensible layouts of warehouses, factories, offices, etc. In a scale
model, physical properties are simply changed in scale and the
relationship of the model to the full-scale system is usually obvious.
However, such simple scale models do have significant
disadvantages. First, a scale model is concrete in form and highly
specific. No one would contemplate using the same scale model for
a chemical factory and a school-the two require distinctly different
buildings. More subtly, to experi- ment with a scale model always
requires physical alteration of the model. This can be tiresome and
expensive. Second, scale models are static. That is, they cannot
show how the various factors interact dynamically. For example,
suppose that a warehouse is being designed. One issue that must be
considered is the relationship between the internal capacity of the
building and the number of loading or unloading bays provided for
vehicles. Though it is easy to design a warehouse that always has
enough internal space-simply make it too big-this is clearly a waste
of money. Given that both the demand for the products and the
produc- tion level will vary, the art is to design a building that
balances the cost of shortages with the cost of over capacity. Such a
balance will vary over time, particularly for seasonal products. No
scale model could consider this. Management scientists tend to
employ mathematical and logical models rather than scale models.
Mathematical models represent the important factors of a system by
a series of equations that may sometimes be solved to produce an
optimal solution. Many of the commonly employed techniques
described in management science textbooks are of this form

8 THE COMPUTER SIMULATION APPROACH

(e.g., mathematical programming, game theory, etc.). For computer

simula- tion, logical models are usually required-though in the case
of system dynamics (see Chapters 13 to 15) these are expressed in
a mathematical form. The simplest way of thinking about logical
models is to consider flow diagrams of various kinds. Industrial
engineers often employ flow process charts in method study (Slack
et aI., 1995) to display the various processes through which
products pass in their manufacture and assembly. That is, the charts
display the logic of the production process. Such a chart might show
that a car body needs to be thoroughly degreased before any
painting can begin. Instead of drawing a chart it is possible to
represent the logic as a set of instructions. If these directions are
clear and unambiguous, then they could be used to show someone
how to do the job. Modern digital computers are logical machines
that obey a sequence of in- structions, thus any sequence of
instructions can form the basis of a computer program-which makes
computer simulation possible. At some stage the simulation model,
which may initially exist on scraps of paper, in agreed documents or
in some formal set of flow diagrams, must be translated into a form
that a computer can recognize and obey. Once in a computable
form, the model may be easily modified so as to permit a wide range
of options to be compared in simulation experiments.

1.4 SIMULATION AS EXPERIMENTATION

Computer simulation involves experimentation on a computer-based
model of some system. The model is used as a vehicle for
experimentation, often in a "trial and error" way to demonstrate the
likely effects of various policies. Those that produce the best results
in the model would be candidates for imple- mentation in the real
system. Figure 1.1 shows the basic idea. Sometimes these
experiments may be quite sophisticated, involving the use of
statistical design techniques. Such sophistication is necessary if there
is a set of different effects that may be produced in the results by
several interacting policies. At the other extreme, the
experimentation may be very simple, taking the form of "what if?"
questions. Thus, if the simulation model

Inp uts (poli ses)

uts Outp cies) SIMULATION (respon

... ......... MODEL(S) ,.

Interaction and experimentation Figure 1.1 Simulation as

experimentation

WHY SIMULATE? 9

represents the financial flows in an organization over the next 12

months, typical questions might be:
. "What if interest rates rise by 3%?" . "What if the market grows by
5% this year?"

To answer these questions, the simulation is carried out with the

appropriate variables of the program set to these values.

1.5 WHY SIMULATE?

Certainly, computer simulation is no panacea. Realistic simulations

may require long computer programs of some complexity. There are
special purpose simulation languages and packaged systems
available to ease this task, but it is still rarely simple. Consequently,
producing useful results from a computer simulation can turn out to
be a surprisingly time-consuming process. In one way, therefore,
computer simulation should be regarded as a last resort-to be used
if all else fails. However, there are certain advantages in employing a
simulation approach in management science and it may be the only
way of tackling some problems. Assuming that a management
scientist does not wish to make an instant "seat of the pants"
judgement of a particular problem, various modes of approach are
possible. First, it may be possible to conduct experiments directly on
the real system. For example, the police may experiment with mock
radar speed traps to see if this reduces the number and severity of
accidents reported. Second, the analyst may be able to construct
and use a mathematical model of the system of interest. For
example, Thomas et al. (2001) describe how mathematical
programming techniques are used in credit scoring and Wright
(1994) describes how heuristics are used to timetable English
cricket. A third possibility is to simulate the system.
1.5.1 Simulation versus direct experimentation

Then why simulate when it will be time consuming and there may be
alterna- tive approaches? Considered against real experimentation,
simulation has the following advantages:

. Cost. Though simulation can be time consuming and therefore

expensive in terms of skilled manpower, real experiments may also
turn out to be expensive-particularly if something goes wrong! .
Time. Admittedly, it takes a significant amount of time to produce
working computer programs for simulation models. However, once
these are written then an attractive opportunity presents itself.
Namely it is possible to simulate weeks, months or even years in
seconds of computer time. Hence a whole range of policies may be
properly compared.

10 THE COMPUTER SIMULATION APPROACH

. Replication. Unfortunately, the real world is rarely kind enough to

allow precise replication of an experiment. One of the skills
employed by physical scientists is the design of experiments that are
repeatable by other scientists. This is rarely possible in management
science. It seems unlikely that an organization's competitors will sit
idly by as a whole variety of pricing policies are attempted in a bid to
find the best. It is even less likely that a military adversary will allow
a replay of a battle. Simula- tions are precisely repeatable. . Safety.
One of the objectives of a simulation study may be to estimate the
effect of extreme conditions, and to do this in real life may be
dangerous or even illegal. An airport authority may take some
persuading to allow a doubling of the flights per day even if they do
wish to know the capacity of the airport. Simulated aircraft cause
little damage when they run out of fuel in the simulated sky. .
Legality. Even when not employed by the mafia there are times
when an analyst may wish to investigate the effect of changes in
legislation. For example, a company may wish to see what the effect
would be on its delivery performance of changes in the laws that
control drivers hours of work.

1.5.2 Simulation versus mathematical modelling

What, then, of the other possibility of building and using a

mathematical model of the system? Here too there are problems.
First, most mathematical models cannot satisfactorily cope with
dynamic or transient effects and operate instead with average
values. However, in any dynamic system, steady state values can be
very misleading, particularly if there is statistical variation in
demand. Even though average demand is met, this may not be true
of peak demand. The challenge is to design such systems to meet
reason- able demand without having idle resources "just in case".
Thus, the model may need to take account of the statistical variation
that is inherent in many systems. Second, though it is debatable
(see Chapter 10) whether this is a good thing, it is possible to
sample from non-standard probability distributions in a simulation
model. However, queuing theory models permit only certain
distributions and therefore cannot cope with many types of problem.
Computer simulation then may well be regarded as the last resort.
Despite this, it is surprising how often such an approach is needed.

1.6 SUMMARY
Computer simulation methods allow experimentation on a computer-
based model of some system. The model is built by carefully
describing the ways in which the system changes state and the rules
that govern its dynamic behaviour. Modelling is best planned on an
incremental and parsimonious basis, with the expectation that the
model will need to be enhanced as

REFERENCES 11

EXERCISES

REFERENCES

knowledge about the system develops. Once built, the model is used
for experimentation, either interactive or classical or both.

(1) Suppose that a public authority is considering various policies for

checking whether goods vehicles are overweight as they arrive at
ferry ports. Discuss whether it might be sensible to consider a
simulation approach. (2) In what types of situation would a
simulation approach be unwise? (3) Spreadsheet packages such as
Microsoft Excel™ are widely used on personal computers. Discuss
what type of simulation these packages allow. (4) If you were the
manager of a factory whose production operations were being
simulated by a management scientist, why might you not be
convinced that the simulation model was valid even if the production
rates output from the simulation were the same as those of your
factory? (5) Consider the reasons why simulation approaches often
form part of the process of designing new manufacturing systems.
(6) Computers are becoming easier to use by non-specialists. Should
managers be encouraged to undertake computer simulations
themselves or is there still a place for the specialist?

Bhaskar R., Lee B.S., Levas A., Petrakian R., Tsai F. and Tulskie W.
(1994) Analyzing a re-engineering business processes using
simulation. Proceed- ings of the 1994 Winter Simulation Conference,
December 1994, Lake Bueno Vista, Florida. Burnett D. and LeBaron
T. (2001) Efficiently modeling warehouse systems. In: B.A. Peters,
J.S. Smith, D.J. Medeiros and M.W. Rohrer (eds), Proceed- ings of
the 2001 Winter Simulation Conference, December 2001, Arlington,
VA. Buxton J.N. and Laski J.G. (1962) Control and simulation
language. The Computer Journal, 5, 3. Ceric V. (1990) Simulation
study of an automated guided-vehicle system in a Yugoslav hospital.
Journal of the Operational Research Society, 41(4), 299-310. Davies
M.N. (1994) Back-office process management in the financial
services-a simulation approach using a model generator. Journal of
the Operational Research Society, 45(12), 1363-73. Davies R.,
Roderick P., Brailsford S.C. and Canning C. (2002). The use of
simulation to evaluate screening policies for diabetic retinopathy.
Diabetic Medicine, 19(9), 763-71.

12 THE COMPUTER SIMULATION APPROACH

De Jong C.D. (2001) Simulating test program methods in

semiconductor assembly test factories. In: B.A. Peters, J.S. Smith,
D.J. Medeiros, and M.W. Rohrer (eds), Proceedings of the 2001
Winter Simulation Conference, December 2004, Arlington, VA.
Dennis S., King B., Hind M. and Robinson S.R. (2000) Applications of
business process simulation and lean techniques in British
Telecommuni- cations pIc. In: K. Kang, J.A. Joines, and R.R. Barton
(eds), Proceedings of the 2000 Winter Simulation Conference,
Orlando, FL. Harrell C.R. (1993) Modeling beverage processing using
discrete event simu- lation. Proceedings of the 1993 Winter
Simulation Conference, December 2000, Los Angeles, CA. Heath W.
(1993) Waterfront capacity-planning simulations. Proceedings of the
1993 Winter Simulation Conference, Los Angeles, CA. Hupert N.,
Mushlin A.I. and Callahan M.A. (2002) Modeling the public health
response to bioterrorism: Using discrete event simulation to design
antibiotic distribution centers. Medical Decision Making, 22(Suppl.),
S17-S25. Jacobson S.H., Sewell E.C. and Weniger W.G. (2001) Using
Monte Carlo simulation to assess the value of combination vaccines
for pediatric immu- nization. In: B.A. Peters, J.S. Smith, D.J.
Medeiros, and M.W. Rohrer (eds), Proceedings of the 2001 Winter
Simulation Conference, December 2001, Arlington, VA. Joustra P.E.
and Van Dijk N.M. (2001) Simulation of check-in at airports. In: B.A.
Peters, J.S. Smith, D.J. Medeiros, and M.W. Rohrer (eds),
Proceedings of the 2001 Winter Simulation Conference, December
2001, Arlington, VA. Ladbrook J. and Januszczak A. (2001) Ford's
power train operations- changing the simulation environment. In:
B.A. Peters, J.S. Smith, D.J. Medeiros, and M.W. Rohrer (eds),
Proceedings of the 2001 Winter Simula- tion Conference, December
2001, Arlington, VA. Lee S., Pritchett A. and Goldsman D. (2001)
Hybrid agent-based simulation for analyzing the national airspace
system. In: B.A. Peters, J.S. Smith, D.J. Medeiros, and M.W. Rohrer
(eds), Proceedings of the 2001 Winter Simu- lation Conference,
December 2001, Arlington, VA. Lu R.F. and Sundaram S. (2002)
Manufacturing process modeling of Boeing 747 moving line
concepts. In: J.M. Charnes, E. Yiicesan and C.-H. Chen (eds),
Proceedings of the 2002 Winter Simulation Conference, December
2001, San Diego, CA. McGuire F. (1994) Using simulation to reduce
lengths of stay in emergency departments. Proceedings of the 1994
Winter Simulation Conference, December 1994, Lake Bueno Vista,
Florida. Miser H.J. and Quade E.S. (eds) (1988) Handbook of
Systems Analysis: Craft Issues and Procedural Choices. John Wiley &
Sons, Chichester, UK. Pidd M. (1987) Simulating automated food
plants. Journal of Operational Research Society, 38(8), 683-92. Pidd
M. (1995) The construction of an object-oriented traffic simulator.
Pro- ceedings of the 3rd BURO Working Group on Transportation,
September 1995, Barcelona, Spain.

REFERENCES 13

Pidd M. (2003) Tools for Thinking: Modelling in Management Science

(2nd edition). John Wiley & Sons, Chichester, UK. Rathi A.K. and
Santiago A.J. (1990) The new NETSIM simulation model. Traffic
Engineering and Control, 31(5), 317. Rivett B.H.P. (1994) The Craft
of Decision Modelling. John Wiley & Sons, Chichester, UK. Robinson
T. (2001) ODIN-An underwater warfare simulation environ- ment. In:
B.A. Peters, J.S. Smith, D.J. Medeiros, and M.W. Rohrer (eds),
Proceedings of the 2001 Winter Simulation Conference, December
2001, Arlington, VA. Salt J. (1991) Tunnel vision. ORIMS Today,
18(1),42-8. Slack N., Chambers S., Harland C. and Johnston R.
(1995) Operations Man- agement. Pitman, London. Thomas L.C.,
Banasik J. and Crook J.N. (2001) Recalibrating scorecards. Journal of
Operational Research Society, 52,981-8. White D. J. (1975) Decision
Methodology. John Wiley & Sons, Chichester, UK. Williams D.L., Finke
D.A., Medeiros D.J. and Traband M. T. (2001) Discrete simulation
development for a proposed shipyard steel processing facility. In:
B.A. Peters, J.S. Smith, D.J. Medeiros, and M.W. Rohrer (eds),
Proceed- ings of the 2001 Winter Simulation Conference, Arlington,
VA. Wright M.B. (1994) Timetabling county cricket fixtures using a
form of tabu search. Journal of the Operational Research Society,
45(7),758-71. Young W., Taylor M.A.P. and Gipps P.G. (1989)
Microcomputers in Traffic Engineering. John Wiley & Sons/Research
Studies Press, Chichester, UK.

A Variety of Modelling Approaches

2.1 GENERAL CONSIDERATIONS

Before producing a dynamic simulation model and thus a computer

program, the analyst must decide on the principal elements of that
model, bearing two things in mind. The first is the nature of the
system being simulated- obviously, the model needs to be a close fit,
a good representation of the system. Needless to say, some
modelling approaches are more suited to certain problems than to
others. The second aspect is the nature of the study being carried
out. That is, what are the objectives of the study, what is the point
of the simulation, what results are expected? Considering both of
these aspects will allow the analyst to decide what level of accuracy
and detail is appropriate for the simulation. There is clearly little
point in producing an extremely detailed simulation if only crude
estimates are required. The practical decisions that need to be made
concern the following, each of which will be considered in this
chapter:

. Time handling. . Stochastic or deterministic durations. . Discrete or

continuous change.
2.2 TIME HANDLING

One advantage of simulation is that the speed at which the

experiment proceeds can be controlled. The essence of a dynamic
simulation is that the state changes of the system are modelled
through time. In management science it is usual to speed up the
passage of time so as to simulate several weeks or months in a few
minutes of computer time. Hence it is important to consider how
time-flow might be handled within the simulation.

2.2.1 Time slicing

Perhaps the simplest way of controlling the flow of time in a

simulation is to move it forward in equal time intervals. This
approach is often described as

16 A VARIETY OF MODELLING APPROACHES

Table 2.1 Job shop order book

Job number
Batch size

Day order expected

1234

200 400 100 200

1 8 14 18

"time slicing" and involves updating and examining the model at

regular intervals. Thus, for a time slice of length dt, the model is
updated at time (t + dt) for changes occurring in the interval (t to (t
+ dt)). One obvious problem with this approach is that some
decision must be taken about the length of the time slice before the
simulation is carried out. For example, the activity levels within a
supertanker terminal may necessitate a time slice of one hour,
whereas for a civil airport the time slice may be more appropriately
set to a half minute or less. Clearly, if the time slice is too large then
the behaviour of the model is much coarser than that of the real
system because it is impossible to simulate some of the state
changes that occur. If, on the other hand, the time slice is too small
then the model is frequently examined unnecessarily (when no state
changes are possible) and this leads to excessively long computer
runs. As a simple example (based on an example in Jones, 1975)*
consider a workshop with just two machines, A and B. Suppose that
the time taken to complete a job on these machines depends on the
size of the job. Thus the job times are:

. Machine A: (batch size/50 + 1) days. . Machine B: (batch size/lOa

+ 3) days.

Suppose too that the workshop only takes on jobs which must be
processed on both machines and that each job must first pass
through machine A as a complete batch and then through machine B
as a complete batch. That is, no batch may be started on either
machine until the previous batch is completed on that machine. If
the workshop expects to receive the four orders shown in Table 2.1.
when will the final batch be complete? The expected job times
(days) are as shown in Table 2.2. Simulating the workshop using a
time slice of one day leads to the times shown in Table 2.3. Thus job
4 is complete at the end of day 32. Following this table through; on
day 1, job 1 arrives and its processing immediately begins on
machine A. Nothing new happens on days 2, 3 or 4 until the end of
day 5 when machine A has finished job 1. Thus on day 6, machine B
starts work on job 1. On day 7 nothing happens. On day 8, job 2
arrives and machine A begins its

* This example is reproduced in revised form with permission from

the Open University from T341 Systems Modelling: Unit 6 Simulation
Modelling. (Q 1975 The Open University Press.

TIME HANDLING 17
Table 2.2 Expected job times

Job number

Machine A

Machine B

1234

5935

5745

Table 2.3 Job shop: time-slicing simulation Jobs queuing Jobs in

progress Jobs queuing Jobs in progress Day For machine A For
machine B Machine A Machine B Day For machine A For machine B
Machine A Machine B 1 17 3 2 2 18 4 3 2 3 19 4 3 2 4 20 3 4 2 5 21
3 4 2 6 22 3 4 2 7 23 3 4 2 8 2 24 4 3 9 2 25 4 3 10 2 26 4 3 11 2
27 4 3 12 2 28 4 13 2 29 4 14 3 2 30 4 15 3 2 31 4 16 3 2 32 4
processing. This is obviously a tedious and inefficient way of
simulating such a simple system, for there is little point in examining
and attempting to update the model each day-on many days,
nothing changes.

2.2.2 Next-event technique

Because many systems include such slack periods of varying length

it is often preferable to use a variable time increment. In this case,
the model is only examined and updated when it is known that a
state change is due. These state changes are usually called events
and, because time is moved from event to event, the approach is
called the next-event technique. Consider again the simple
workshop. Table 2.4 shows the results of a next- event simulation of
this system. Notice that the table is much smaller than that required
for a time-slicing approach. The method focuses on the progress of
each job as it passes through the workshop. The events are:

. A job arrives. . Machine A starts a job.

18 A VARIETY OF MODELLING APPROACHES

Table 2.4 Job-shop: next-event simulation Machine A Machine B Job

No. Arrival date Start Finish S ta rt Finish 1 1 1 5 6 10 2 8 8 16 17 23
3 14 17 19 24 27 4 16 20 24 28 32
. Machine A finishes a job. . Machine B starts a job. . Machine B
finishes a job.

Each of these events may occur a maximum of four times during the
simula- tion, once for each job. In fact, as Table 2.4 shows, some of
these coincide and the model need only be updated on 16 occasions.
By way of contrast, Table 2.3 shows the inevitable 32 updates of a
time-slicing approach. Though the halving of the number of state
changes is only a feature of this example, a next-event model will
usually require fewer state changes than one employing time slicing.

2.2.3 Time slicing or next event?

Thus, a next-event technique has two advantages over a time-slicing

approach. The first is that the time increment automatically adjusts
to periods of high and low activity, thus avoiding wasteful and
unnecessary checking of the state of the model. The second is that it
makes clear when significant events have occurred in the simulation.
On the other hand, the simulation software that drives the
simulation must be more intelligent, since it must manage a diary of
future events. This means that it must hold more information than is
needed for a time-slicing simulation. However, unless the actual
events of a system do occur at regular intervals, a next- event
technique is usually better. Of course, some systems have events
that occur at regular intervals. For example, a superstore may check
its stock levels at the same time each day and replenishment may
similarly arrive at predictable times. In such cases it is quite
adequate to update the model at regular intervals to allow for the
intervening changes. Nevertheless, it should be noted that the next-
event technique is more general since, if the events in a system
occur at regular intervals (once per day, perhaps), the next-event
technique will act as if it were a time-slicing approach. The reverse is
not true.

2.3 STOCHASTIC OR DETERMINISTIC?

A system is deterministic if its behaviour is entirely predictable.

Provided that the system is perfectly understood, then it is possible
to predict precisely what will happen. A cycle of operations on an
automatic machine may be determi-

STOCHASTIC OR DETERMINISTIC? 19

nistic in this sense. Each repeated identical cycle will take the same
length of time unless the conditions influencing the cycle times are
altered. A system is stochastic if its behaviour cannot be entirely
predicted, though some statement may be made about how likely
certain events are to occur. For example, a lecturer may give the
same lecture to several sets of students but the duration of the
lecture may vary from occasion to occasion. Statistical statements
may be made about the duration of the lecture: for example, that it
is normally distributed with a mean of 50 minutes and a standard
deviation of 3 minutes. Thus it is highly likely that the duration of
the lecture will exceed 48 minutes. However, it is impossible to
precisely state how long a par- ticular delivery of the lecture will last
unless the lecturer's behaviour can be completely controlled-and that
of the class too! In some senses, the distinction between stochastic
and deterministic systems is artificial. It is more a statement of the
amount of knowledge about a system or the amount of control over
that system exercised by an observer. However, it is important to
notice that both stochastic and deterministic simulations are
possible.

2.3.1 Deterministic simulation: a time-slicing example

Any deterministic simulation model contains no stochastic elements

and a simple example was the four-job simulation of the workshop
of Section 2.2. As another example, this time one which can be
formulated as a set of differ- ence equations, consider the case of
Big Al and his recruitment problems. After his release from gaol
(jail), Big AI, a well known gangster, decides to rebuild his mob for
more assaults on the banks of Bailrigg County. This time he plans a
large-scale operation and reckons that he would like to have 50
mobsters working for him within six months. He currently has none.
His previous experience in forming a mob suggests that he can
recruit at a weekly rate equal to one quarter of the difference
between his ideal mob size (50) and the number currently in the
mob. His problem is that mobsters are caught by the cops with
depressing frequency. Indeed, Happy Harry, chief of the Bailrigg
County cops, boasts that his men will catch 5% of Big AI's active
mobsters.in each week and they receive gaol sentences of at least
12 months each. Fortunately, 10% of those in gaol escape each
week and rejoin Big AI's mob. Big Al himself has other ways to
satisfy the needs of the local police and does not expect to be
arrested again. How large will his mob size be after 10 weeks? One
approach to Big AI's problem is to use a simple time-sliced
simulation based on a two-part set of difference equations. To do
this requires some variables to be defined.

Variables
Suppose that any 2-week interval can be represented as starting at
time t - 1 (the first weekend), with an intervening weekend at time t
and a final weekend at time t + 1. Variables of two types may now
be defined.

20 A VARIETY OF MODELLING APPROACHES

(1) Aggregated values at definite time points Consider the time point
t Mob size = MS t N umber in gaol = NG t (2) Variables which
represent rates which are constant over an interval Consider the
interval t - 1 to t

AI's recruitment rate = RECt-l,t The rate at which gangsters are

arrested = ARRt-l,t The rate at which gangsters escape from gaol =
ESCt-l,t The target mob size is a constant, TARGET

Hence, the following equations can be formulated:

(1) Aggregated values at time t MS t = MS t - l + (RECt-l,t - ARRt-l,t)

+ ESCt-l,t NG t = NG t - l + (ARRt-l,t - ESCt-l,t) That is, the value of
MS at time t is the value of MS at time t - 1, plus the changes that
occur over the interval t - 1 to t. The latter is number of recruits,
plus the number of escapees, minus the number of mobsters
arrested over that interval.
(2) Constant rates over the next week RECt,t+l = (TARGET - MS t
)/4 ARRt,t+l = MS t *0.05 ESCt,t+l = NGd10

The system may now be simulated using a simple spreadsheet in

which the columns represent the different variables and the rows
represent the time points (weekends). The four constants, TARGET
and the three numerical parameters, can be placed as values on the
spreadsheet to make experimenta- tion rather easier. The number in
gaol and the mob size need to be specified for time point 0 (just
before the start of week 1) and then the normal spreadsheet-linked
computations can be employed. To check the computations by hand,
do the following. Starting with time point t = 0:

(1) Write down the values of the mob size (MS) and number in gaol
(NG). (2) Write down the rates (recruitment, arrest, escape) over
the next interval.

Move time to the next point (t = 1) and repeat; continue with t =

2,3, . . . ,10. The results of the simulation for a 10-week period are
shown in Table 2.5 and in Figure 2.1. Table 2.5 is slightly deceptive,
for it hides the fact that the rates are actually computed over
intervals of length dt, which is 1 week in this case, rather than at the
time points 0, 1, 2, etc. Strictly speaking, these rate values should
be printed in alternate lines, between the other values. It is

STOCHASTIC OR DETERMINISTIC? 21
Table 2.5 Big AI's recruitment problem Week Recru it rate Arrest rate
Escape rate No. in gaol Mob size 0 0.00 0.00 1 12.50 0.00 0.00 0.00
12.50 2 9.38 0.63 0.00 0.63 21.25 3 7.19 1.06 0.06 1.63 27.44 4
5.64 1.37 0.16 2.83 31.87 5 4.53 1.59 0.28 4.14 35.09 6 3.73 1.75
0.41 5.48 37.48 7 3.13 1.87 0.55 6.81 39.28 8 2.68 1.96 0.68 8.09
40.68 9 2.33 2.03 0.81 9.32 41 .78 10 2.05 2.09 0.93 10.48 42.68
50 40 (]) N "en 30 .D 0

20 10 0 0 2 3 4 5 6 7 8 9 10 Week Figure 2.1 Big AI's mob size

clear that Big Al does not reach his target of 50 mobsters within ten
weeks. Indeed, he may have trouble over a much longer period due
to the fact that the recruitment rate is a function of the gap between
the target and actual mob sizes. To get to 50 he may need to find
other incentives. There is one obvious problem about this model. It
uses real values but even gangsters prefer to be treated as integers.

2.3.2 Stochastic simulation

Many systems behave stochastically and must therefore be

simulated by a model with stochastic elements. This means that
probability distributions are used in such stochastic simulation
models. As the simulation proceeds, samples are taken from these
distributions so as to mimic the stochastic behaviour. As an example,
consider the following replacement problem. A multi-user computer
system includes two disk units which, being mech- anical. are prone
to failure. If a disk unit fails in service, users lose their files

22 A VARIETY OF MODELLING APPROACHES

Table 2.6 Probability of disk unit failure

Days since repair or maintenance

Probability of failure

1 2 3 4 5 6 >6

0.05 0.15 0.20 0.30 0.20 0.10 0.00

(and their tempers) which need to be restored. Restoration is

achieved by copying on to the disk back-up copies of the files held
on magnetic tapes. This restoration is inconvenient and so a new
operating policy is being considered. At the moment, the disk units
are repaired and restored as and when they fail. The proposal is to
introduce a joint repair system. Table 2.6 shows the probability of a
disk unit failing in the days following its last repair. That is, 5% of
the units are expected to fail 1 day after repair or maintenance, 15%
after 2 days, etc. Under the current repair policy, it costs $50 per
disk to repair and restore a failed unit. The joint repair system would
operate as follows. When either unit fails, the failed unit is repaired
and restored at a cost of $50 per unit. If operational. the other unit
will be cleaned at a cost of $25. Cleaning a disk places it in a state
equivalent to having been just repaired and restored. Is the new
joint repair system cost effective? This question can be answered by
a simple stochastic simulation which involves random sampling from
the failure distribution of the disks given earlier. Details of sampling
methods are given in Chapter 10, but for present purposes a simple
method can be used. Figure 2.2 shows a histogram of the disk
failure distribution. In Figure 2.3, the data have been rearranged to
show the cumulative probability of the various lives. For example,
the probability of a disk lasting up to and including 3 days is 0.40

0.30 0.25 g 0.20 :3 co .D 0.15 0 C: 0.10 0.05

0.00

Figure 2.2

3 4 life (days) Histogram of failure probabilities

STOCHASTIC OR DETERMINISTIC? 23

1.0

g 0.8 :3 co .Q e 0.6 c.. Q) >

0.4 "S E :J 0.2 ()

0.0

Figure 2.3

3 4 Life (days) Histogram of cumulative failure probabilities

(0.05 + 0.15 + 0.20). Using the cumulative form of Figure 2.3 and
random number tables, random samples can be taken from the life
distribution by asso- ciating a life (in days) with each random
number. An extract from a random number table is shown in Table
2.7, the values range from 00 to 99, and any number in that range
has an equal probability of appearing at any position in the table. If
these random numbers are divided by 100, so that their range is
0.00 to 0.99, then Figure 2.3 may be used to generate random
samples as follows. The first random number in the table is 27 (Le.,
0.27); if this is marked on the vertical axis of Figure 2.3, then the
corresponding point on the horizontal axis is 3 days. That is, 3 days
is the life associated with the random number 0.27. In general. each
life may be asso- ciated with a range of random numbers as shown
in Table 2.8. Table 2.9 shows a 50-day simulation of the two policies.
In the case of the separate replacement policy, replacements of each
of the two units have been

Table 2.7 Some random numbers 27 62 36 30 57 78 22 02 89 22 04

97 43 30 45 12 03 87 16 50 92 26 00 82 58 10 78 44 55 05 21 50
49 83 49 39 25 81 03 99 77 71 43 06 90 09 04 97 07 64 40 39 69
42 63 80 07 85 65 70 60 57 42 97 29 92 84 54 66 91 34 10 78 81
97 99 08 19 15 63 35 37 13 56 88 09 36 40 07 55 04 24 69 52 44
14 61 59 31 50 24 26 29 31 57 17 38 44 03 29 26 63 00 44 64 09
93 15 52 35 91 37 65 32 84 37 80 94 48 46 23 52 10 77 27 40 34
13 73 53 55 89 99 78 50 11 43 43 54 16

24 A VARIETY OF MODELLING APPROACHES

Table 2.8 Look-up table linking random numbers and disk life
Life (days) Associated random numbers

1 0.00-0.04 2 0.05-0.19 3 0.20-0.39 4 0.40-0.69 5 0.70-0.89 6 0.90-

0.99

Table 2.9 Next-event simulation of the disk repair policies Separate

repair Joint repair Unit A Unit B Random number Life Time of failure
Random number Life Time of failure Time of failure 1 0.27 3 3 0.24
3 3 3 2 0.62 4 7 0.26 3 6 6 3 0.36 3 10 0.29 3 9 9 4 0.30 3 13 0.31
3 12 12 5 0.57 4 17 0.57 4 16 16 6 0.04 1 18 0.26 3 19 17 7 0.97 6
24 0.63 4 23 21 8 0.43 4 28 0.00 1 24 22 9 0.30 3 31 0.44 4 28 25
10 0.45 4 35 0.64 4 32 29 11 0.92 6 41 0.91 6 38 35 12 0.26 3 44
0.37 3 41 38 13 0.00 1 45 0.65 4 45 39 14 0.82 5 50 0.32 3 48 42
15 0.58 4 0.84 5 53 46 16 0.21 3 0.23 3 49 17 0.50 4 0.52 4 53 18
0.49 4 0.10 2 19 0.83 5 0.77 5 20 0.49 4 0.27 3

simulated until the failure time (the cumulative life) of each is

greater than or equal to 50 days. For this policy, unit A needed 14
repairs and restores; unit B needed 15. Hence, 29 units were used at
a cost of $50 per unit, giving a total cost of $1450 over 50 days.
Exactly the same random numbers are used to simulate the second
policy, which is also shown in Table 2.9. In this case, the units are
considered in

DISCRETE OR CONTINUOUS CHANGE? 25

pairs. Hence, starting at time zero, the first A unit would last for 3
days if allowed to do so, as would the first B unit. In this case,
therefore the first re- placement takes place after 3 days. The
second pair would last 4 days (in the case of A) and 3 days (in the
case of B). Hence the second replacement will take place 3 days
after the first (Le., after 6 days). The simulation continues in this
way until the replacement time (the joint cumulative life) is at or
greater than 50 days. This shows that there were 17 such joint
repairs in the period and that, of these 17, 9 involved both units in
repair and restore (Le., they both failed at the same time-
coincidences). The other 8 of the 17 involved one unit in a repair
and restore, whilst the other need only be cleaned. Hence, there
were 26 repairs and restores at $50 per unit and 8 clean-ups at $25
per unit. Giving a total cost of $1500. Thus on the basis of a single
simulation, the new policy costs $50 more over a 50-day period.
However, it would be wrong to assume that a separate repair policy
is therefore more cost effective. If a different set of random numbers
were used, the result could have been different for both policies and
the new policy might appear cheaper. Chapter 4 returns again to this
issue in its discussion of static Monte Carlo simulations. In stochastic
simulations, it is important to realize that the results are dependent
on the samples taken. Hence it is usual to make several simulation
runs each with a distinct sampling pattern before drawing any
conclusions. Chapter 11 gives details. The sampling methods
employed in stochastic simu- lations are well developed and
documented. Commonly used methods are given in Chapter 10 and
most simulation software systems have appropriate subroutines or
procedures ready for use. Thus distribution sampling should not be a
problem in model building. More likely, controlling the sequence of
events is the major problem in model building. The interaction of the
entities is responsible for the sequence of events, and in the above
example these are trivial. But for complex systems a sensible
structure is needed. Various common approaches are described in
Chapters 5 and 6.
2.4 DISCRETE OR CONTINUOUS CHANCE?

In the job shop simulation, it was convenient to regard the system

as moving from state to state through time. The concern with
individual batches was whether their machining was complete or
not, rather than with the rate at which the machining was
proceeding. The variables included in a simulation model can be
thought of as changing value in four ways:

(1) Continuously at any point of time. Thus, the values are changing
smoothly and not discretely and the values taken are accessible at
any time point within the simulation. (2) Continuously but only at
discrete time points. In this mode, the values again change smoothly
but can only be accessed at predetermined times. (3) Discretely at
any point of time. In these simulations, state changes are easily
identifiable but can occur at any point of time.

26 A VARIETY OF MODELLING APPROACHES

(4) Discretely and only at discrete points of time. The state changes
can only occur at specified points of time.

Historically, computer simulation applications have tended to divide

into those employing discrete change and those that allow the
variables to con- tinuously change value.
2.4.1 Discrete change

Consider an underground railway in which trains move from station

to station, picking up and depositing passengers at each. Viewed
from the per- spective of discrete change there are a number of
obvious system events. For example:

. Train stops at station. . Doors now open. . Doors now closed. .

Train starts to leave station.

Thus to simulate this system using a discrete model, the time taken
to travel between stations or to open the doors would either be
known deterministically or could be sampled from some appropriate
distribution. Thus, for example, when the train starts to leave a
station its arrival at the next station could be scheduled by referring
to this "known" journey time. In a discrete simulation the variables
are only of interest as and when they point to a change in the state
of the system. Chapters 5 to 7 are devoted to the exposition of
methods suitable for discrete simulation.

2.4.2 Continuous change

If the underground railway were to be simulated via a model that

allowed con- tinuous change, then the variables would be
continuously changing their values as the simulation proceeds.
Consider, for example, the train as it travels between stations. If the
locomotive is electrically powered, its speed will increase smoothly
from rest until it reaches an appropriate cruising rate. The speed
does not change by discrete amounts. Thus, if the results of the
simulation are to include the state of the system in relation to the
continuous variable "speed", then a continuous change model is
needed. These continu- ous changes could be represented by
differential equations which would, in theory, allow the variables to
be computed at any point in time. In considering continuous change
models it must be recognized that digital computers operate only
with discrete quantities. Hence, changes cannot actually be
occurring continuously within a "continuous" simulation on a typical
computer. In system dynamics (see Chapters 13 to 15) the apparent
continuity is achieved by allowing the variables to be inspected or
changed at a multitude of fixed points in simulated time. This
provides an approximation to continuous change, which is often
good enough.

EXERCISES 27

Continuous simulation "proper" is not covered in detail in this book

because management scientists seem to be more often concerned
with systems that can satisfactorily be simulated discretely. More
often continuous simulations are the concern of economists in
modelling the behaviour of economic systems via sets of differential
equations, or the task of engineers designing equipment. Early
continuous simulations were mainly carried out using analogue
computers. Though these have some appeal for those with an
interest in electrical hardware, they tend to be tedious to reprogram
and of limited accuracy. Therefore, analogue digital simulators, such
as CSMP (IBM, 1970), were developed. Early versions ofthese
simulators employed the block diagram terminology of analogue
computers, a description of the block diagram being the "program"
from which a digital computer could simulate an analogue computer.
Later versions allow systems to be directly represented as sets of
differential equations that are integrated numerically.

2.4.3 A few words on simulation software

Chapter 8 introduces two commonly used dynamic simulation

packages, Micro Saint (Micro Analysis and Design, 2003) and
SIMUL8 (Simul8 Corpora- tion, 2003). Both of these packages are
based on a next-event simulation technique and both provide easy-
to-use sampling routines that support full stochastic simulation with
a minimal requirement for computer program- ming. Though it is
possible to "bend" these packages for continuous simula- tion, this is
probably unwise. System dynamics simulations (see Chapters 13 to
15) are usually based on time slicing and this is evident in packages
such as iThinkjStella (High Performance Systems, 2003), Powersim
(Powersim Corporation, 2003) and Vensim (Ventana Systems, 2003).
Though system dynamics packages do provide some stochastic
sampling routines, these are best thought of as add- ins that are
rarely used. The vendors of simulation software systems realize that
the separation of discrete and continuous simulation is somewhat
artificial. Consequently, a number of simulation software systems
allow the user to program discrete, continuous or mixed models.
Recent examples include GoldSim (GoldSim Technology Group,
2003), AweSim (FrontStep APS, 2003) and Extend (Imagine That,
2003).

EXERCISES
(1) Use a spreadsheet program to simulate Big AI's problem as
described in Section 2.3.1. (2) Place the constants as values on the
spreadsheet and vary these so as to carry out experiments. Use the
graphing facilities to see how the mob size builds up. (3) Try using
your spreadsheet to model the disk failure problem. This is
straightforward in the time-slicing case, but slightly more difficult in
the next-event case.

28 A VARIETY OF MODELLING APPROACHES

REFERENCES

(4) Check out your spreadsheet program to see what control, if any,
it gives you over the random numbers which are produced. If it does
give you proper control, then run the disk failure problem 10 times
and try to understand the variation in the results. (5) If you know a
computer programming language, write a program to simulate the
disk failure problem. Try the time-slicing case first, then the next-
event case. (6) Discuss why analogue computers are almost never
used in management science despite the fact that many systems do
change continuously.

FrontStep APS (2003) http://www.pritsker.com/awesim.htm GoldSim

Technology Group (2003) http://www.goldsim.com/ High
Performance Systems (2003) http://www.hps-inc.com/ IBM
Corporation (1970) Introduction to 1130 Systems Modelling Program
II (CSMP II, GH200848-1). IBM, White Plains, New York. Jones L.
(1975) Simulation Modelling (Unit 6, Course T341, Systems
Modelling). Open University Press, Milton Keynes, UK. Imagine That
 Another Random Document on
Scribd Without Any Related Topics
authors of the detested interim, published, in A.D. 1555, his
Propositiones de libero arbitrio, in defence of Melanchthon’s
synergism. The leaders of the Gnesio-Lutherans, Arnsdorf in
Eisenach, Flacius in Jena, and Musacus in Weimar, felt that
they durst not remain silent, and so they maintained, as alone
the genuine Lutheran doctrine, that the natural man cannot
co-operate with the workings of Divine grace upon him, but
can only oppose them. By order of the Duke John Frederick
they prepared at Weimar, in A.D. 1559, as a new manifesto of
the restored Lutheranism, a treatise containing a refutation of
all the heresies that had hitherto cropped up within the
Lutheran church. One of those invited to take part in the
work, Victorin Strigel, professor at Jena, was made to suffer
for the sympathy which he evinced for synergism by enduring
close and severe imprisonment. The duke, however, soon
again became more favourable to Strigel, who in A.D. 1560
vindicated himself at a public disputation in Weimar against
Flacius, and was soon afterwards called to Leipzig. When in
A.D. 1561 the duke set up a consistory in Weimar, and
transferred to it the right hitherto exclusively exercised in
Jena of ecclesiastical excommunication and the censorship of
theological books, and the Flacian party opposed this
“Cæsaro-papism” with unmeasured violence, all the
adherents of the party were driven out of Jena and out of the
whole territory, and their places filled with Melanchthonians.
This victory of Philippism, however, was of but short duration.
In order to regain the lost electoral rank, the duke allowed
himself to be beguiled into taking part in the so-called
Grumbach affair. He was cast into the imperial prison, and his
brother John William, who now assumed the government,
hastened, in A.D. 1567, to restore the overthrown theological
party. Even in electoral Saxony interest in the Catholicizing
synergism, at least, after Melanchthon’s death, in A.D. 1560,
was gradually lost sight of in proportion as the controversy
about the Calvinistic doctrine of the Lord’s Supper gradually
gained prominence.
§ 141.8. The Flacian Controversy about Original Sin,
A.D. 1560-1575.―In the heat of the controversy with Strigel at
the conference at Weimar, in A.D. 1560, Flacius had committed
himself to the statement that original sin in man is not
something accidental, but something substantial. His own
friends now urged him to retract this proposition, which his
opponents had branded as Manichæan. Its author had not
indeed intended it in the bad sense which it might be
supposed to bear. Flacius, however, was of a character too
dogged and obstinate to agree to recall what he had uttered.
Expelled with the rest of the Lutherans in A.D. 1562, and not
recalled with them in A.D. 1567, he wandered without any
fixed place of abode, driven away from almost every place
that he entered, until shortly before his death he recalled his
overhasty expression. He died in the hospital at Frankfort-on-
the-Maine, in A.D. 1575. In him a powerful character and an
amazing wealth of learning were utterly lost in consequence
of unpropitious circumstances, which were partly his fault and
partly his misfortune.

§ 141.9. The Lutheran Doctrine of the Lord’s

Supper.―The union effected by the Wittenberg Concord of
A.D. 1536 (§ 133, 8) with the South German cities, which
originally favoured Zwinglian views, had been in many cases
threatening to dissolve again, and the attacks of the men of
Zürich obliged Luther in A.D. 1544 to compose his last
“Confession of the Holy Sacrament against the Fanatics.” The
breach with the Zwinglians was now seen to be irreparable,
but it appeared as if it were yet possible to come to an
understanding with the more profound theory of the Lord’s
Supper set forth by Calvin. To carry out this union was a
thought very dear to the heart of Melanchthon. He had the
conviction, not indeed that the Lutheran doctrine of the real
presence of the body and blood in the bread and wine is
erroneous, but rather that by the Calvinistic doctrine of a
spiritual enjoyment of the body and blood of Christ in the
supper by means of faith no essential element of religious
truth was lost, and so he sought thereby to get over the
difference in confession and doctrine. But with this
explanation the strict Lutherans were by no means satisfied,
and long continued and extremely passionate discussions
were carried on in the various Lutheran countries, especially
in Lower Saxony, in the Palatinate, and in the electorate. But
the controversy was not restricted to the question of the
supper; it rather went back upon a deeper foundation. Luther,
carrying out the principles of the third and fourth œcumenical
councils, had taught that the personal connection of the two
natures in Christ implies a communication of the attributes of
the one to the other, communicatio idiomatum, that therefore
Christ, since He has by His ascension entered again upon the
full exercise of His attributes, is, as God-Man, even in respect
of His body, omnipresent, ubiquitas corporis Christi, and
refused to allow himself to be perplexed by the
incomprehensibility for the human understanding of an
omnipresent body. It is here that we come upon the radical
distinction between Luther’s view and that of Zwingli and
Calvin, according to which the body of Christ cannot be at
one and the same time in heaven at God’s right hand and on
the earth in bread and wine. But Calvin, as well as Zwingli,
from his very intellectual constitution, could only regard the
Lutheran doctrine of the ubiquity of the glorified body of
Christ as an utter absurdity, and so, repudiating the
communicatio idiomatum, he taught that the glorification of
Christ’s body is restricted to its transfiguration, and that now
in heaven, as before upon the earth, it can be present only in
one place. A necessary consequence of this view was the
rejection of His corporeal presence in the supper, and at the
very most the admission of a communication in the sacrament
to believers of a spiritual influence from the glorified body of
Christ.―The ablest vindicator of the Lutheran doctrine of the
supper in this aspect of its development was the
Württemberg reformer John Brenz (§ 133, 3). In the
Syngramma Suevicum of A.D. 1525 (§ 131, 1), he has taken
his place most decidedly on the side of Luther, and this he
had also done again, in A.D. 1529, at the Marburg Conference
(§ 132, 4). Then in A.D. 1559, as provost in Stuttgart, in
consequence of the doubtful attitude of a Swabian pastor on
the question of the supper, he summoned a synod at
Stuttgart, before which he laid a confession which expressed
the doctrine of the supper and the ubiquity in strict
accordance with Lutheran views. In defence of the idea of
ubiquity he quoted Ephesians iv. 10, as affording sufficient
Scripture support. The synod unanimously adopted it, and the
duke gave approval to this Confessio et doctr. theologor. et
ministror. Verbi Dei in Ducatu Wirtb. de vera præsentia Corp.
et sang, J. Chr. in Cœna Domini, by ordering that all
preachers should adopt it, and that it should have symbolic
authority throughout the Württemberg church. Melanchthon,
who had hitherto been on particularly intimate terms with
Brenz, was very indignant at this “unseasonable” creed-
making in “barbarous Latin.” Brenz, however, would not be
deterred from giving more adequate expression and
development to the objectionable dogma, and for this
purpose published, in A.D. 1560, his book, De personali unione
duarum natur. in Christo.

§ 141.10. Cryptocalvinism in its First Stage, A.D. 1552-

1574.―The struggle of the Gnesio-Lutherans against Calvin’s
doctrine of the supper, and the secret favour shown toward it
by several Lutheran theologians, was begun in A.D. 1552 by
Joachim Westphal, pastor in Hamburg. Calvin and Bullinger
were not slow in giving him a sharp rejoinder. In a yet more
violent form the dispute broke out in Bremen, where the
cathedral preacher Hardenberg, and in Heidelberg, where the
deacon Klebitz, entered the lists against the Lutheran dogma.
In both cases the struggle ended in the defeat of Lutheranism
(§ 144, 1, 2). In Wittenberg, too, the Philippists George Major,
Paul Eber, Paul Crell, etc., supported by the very influential
court physician of the electoral court of Saxony, Caspar
Peucer, Melanchthon’s son-in-law, from A.D. 1559 successfully
advanced the interests of Cryptocalvinism. Melanchthon
himself, however, was not to live to see the troubles that
arose over this, a truly gracious dispensation of Providence on
behalf of a man already sorely borne down and trembling
with hypochondriac fears, to have him thus delivered a rabie
theologicorum. He died on 19th April, A.D. 1560. While the
Elector Augustus, A.D. 1553-1586, intended that his
Wittenberg should always be the main stronghold of strict
Lutheranism, the Philippists were always coming forward with
more and more boldness, and sought to prepare the way for
themselves by getting all places filled with members of their
party. They persuaded the elector to give a nominative
authority throughout Saxony to a collection of Melanchthonian
doctrinal and confessional documents compiled by them,
Corpus doctrinæ Philippicum s. Misnicum, 1560. The
Wittenberg Catechism, Catechesis, etc., ad usum scholar.
puerilium, 1571, set forth a doctrine of the sacraments and
the person of Christ so manifestly Calvinistic, that even the
elector was obliged to give way on account of the strong
objections brought against it. The Philippists, however,
succeeded in satisfying him by the Consensus Dresdensis, of
10th Oct., A.D. 1571, to this extent, that after the death of
Duke John William, in the exercise of his authority as regent,
he was induced to expel the Lutheran zealots Wigand and
Hesshus from Jena, and in A.D. 1573 had more than a
hundred clergymen of the duchy of Saxony deposed. In
Breslau their interests were also zealously advanced by the
influential imperial physician, John Krafft, to whom the
Emperor Maximilian II. had granted a patent of nobility in
A.D. 1568, with the new name of Crato von Crafftheim.
Another Silesian physician, Joachim Curæus, also a scholar of
Melanchthon, published in A.D. 1574, without any indication of
author’s name, place of publication, or date of issue, his
Exegesis perspicua controversiæ de cœna, which represented
Melanchthon’s doctrine of the Lord’s Supper as the only
tenable one, controverted that of the Lutherans as popish,
eulogized that of the Reformed church as one most honouring
to God, and urgently counselled union with the Calvinists. The
warm recommendation of this treatise on the part of the
Wittenberg Philippists, however, rather contributed to its
failure. For now, at last, even the elector had become
convinced of the danger that threatened Lutheranism through
hints given him by the princes, and information obtained from
intercepted letters. The Philippists were banished, their chiefs
thrown into prison, Peucer being confined for twelve years,
A.D. 1574-1586. A thanksgiving service in all the churches and
memorial medal celebrated the rooting out in A.D. 1574 of
Calvinism, and the final victory of restored Lutheranism.―In
Denmark, Nicholas Hemming, pastor and professor at
Copenhagen, distinguished alike by adequate scholarship and
rich literary activity, and by mildness and temperateness of
character, and hence designated the Preceptor of Denmark,
was the recognised head of the Melanchthonian school. As a
decided opponent of the doctrine of ubiquity, though
otherwise on all points, and especially in his doctrine of the
Lord’s Supper, a good Lutheran, he fell under the suspicion of
the German Gnesio-Lutherans as a Cryptocalvinist, and was
accordingly opposed by them. In A.D. 1579, by order of the
Elector Augustus, his brother-in-law, the King of Denmark
removed him from his offices in Copenhagen, appointing him
to a canonry in the cathedral at Roeskilde, where in A.D. 1600
he died.

§ 141.11. The Frankfort Compact, A.D. 1558, and the

Naumburg Assembly of Princes, A.D. 1561.―After the
disgraceful issue of the Worms Conference of A.D. 1557
(§ 137, 6), the Protestant princes, the electors Augustus of
Saxony, Joachim of Brandenburg, and Ottheinrich of the
Palatinate, with Philip of Hesse, Christopher of Württemberg,
and the Count-palatine Wolfgang, who were gathered
together about the Emperor Ferdinand, consulted as to the
means which they should employ to insure and confirm the
threatened unity of the evangelical church of Germany. The
result of their deliberations was, that they agreed to sign a
statement drawn up by Melanchthon and known by the name
of the Frankfort Compact, in which they declared anew
their unanimous attachment to the doctrine set forth in the
Augustana, the Variata, and the Saxonica (§ 136, 8), and in
regard to controversial questions that had been discussed
within the church expressed themselves in moderate terms as
inclined to the views of Melanchthon. The Flacian party in
Jena hastened to set forth their opposing sentiments in the
manifesto of A.D. 1559, already referred to, in which the strict
Gnesio-Lutheranism was laid down in the hardest and boldest
manner possible.―The divisions that arose within the
Lutheran church after Melanchthon’s death and the imminent
reassembling of the Tridentine Council led the evangelical
princes of Germany, who, with the exception of Philip of
Hesse, all belonged to a new generation, once more to put
forth every effort to restore unity by adoption of a common
evangelical confession. At the Assembly of Princes
appointed to meet for this purpose at Naumburg in
A.D. 1561, most of them appeared personally. There was no
thought of preparing a new confession, because it was feared
that in those times of agitation it might be impossible to draw
up such a document, or that, even if they succeeded in doing
so, it might not close the breach, but rather widen it. Thus
the only alternative remaining was to attempt the healing of
the schism by reverting to the standpoint of the Augsburg
Confession. But then the question arose whether the original
form of statement of A.D. 1530, or its later elaboration of
A.D. 1540, should be taken as the basis of union
negotiations.―This at least was to be said in favour of the
latter, that it had been unanimously adopted as the common
confession of all the evangelicals of Germany at the peace
Conference of Worms in A.D. 1540, where even Calvin had
signed it, and at Regensburg in A.D. 1541 (§ 135, 2, 3); and
now Philip of Hesse and Frederick III. of the Palatinate came
forward decidedly in its favour. But all the more persistently
did the Duke John Frederick of Saxony oppose it, and make
every endeavour to get the rest of the princes to give their
votes in favour of the Augsburg Confession of A.D. 1530. But
the duke’s further wish to have added to it the Schmalcald
Articles found very little favour. Finally a compromise was
effected, in accordance with which, in a newly drawn up
preface, the Apology of the Augustana, as well as the edition
of A.D. 1540, was acknowledged, while the Schmalcald
Articles, as well as the Confessio Saxonica (§ 136, 8) and the
Frankfort Compact, were passed over in silence. John
Frederick now demanded the adoption of an express
condemnation of the Calvinising Sacramentarians. This led to
a hot discussion between him and his father-in-law, the
elector-palatine. He took his departure on the following day
without having received his dismissal, leaving behind him a
sharply worded protest. Ulrich of Mecklenburg also refused to
subscribe, but allowed himself at last to be persuaded into
doing so. At the sixteenth session two papal legates
personally delivered to the princes a brief inviting them to
attend the council. This latter, however, was returned
unopened when they discovered in the address the usual but
artfully concealed formula “dilecto filio.” Also the demand of
the imperial embassy accompanying the legates to take part
in the council was determinedly rejected, because that would
mean not revision but simply a continuation of the previous
sessions of the council, at which the evangelical doctrine had
already been definitely condemned.

§ 141.12. The Formula of Concord, A.D. 1577.―Already

for a long time had the learned chancellor Jac. Andreä of
Tübingen wrought unweariedly for the restoration of peace
among the theologians of the Lutheran church. In order also
to win over the general membership in favour of peace, he
attempted in six popular discourses, delivered in A.D. 1573, to
instruct them in reference to the points in dispute and proper
means for overcoming these differences. He was so
successful in his efforts, that he soon ventured to propose
that these lectures should be made the basis of further
negotiations. But when Martin Chemnitz, the most
distinguished theologian of his age, pronounced them
unsuitable for that purpose, Andreä wrought them up anew in
accordance with Chemnitz’s critical suggestions into the so
called “Swabian Concord.” But even in this form they did not
satisfy the theologians of Lower Saxony. The Swabian
theologians, however, in their criticisms and emendations, had
answered various statements in it, and in A.D. 1576 they
produced a new union scheme, drafted by Luc. Osiander,
called the “Maulbronn Formula.” The Elector Augustus of
Saxony then summoned a theological convention at Torgau,
at which, besides Andreä and Chemnitz, there were also
present Chytræus from Rostock, as well as Körner and Andr.
Musculus from Frankfort-on-the-Oder. They wrought up the
material thus accumulated before them into the “Book of
Torgau,” of A.D. 1576. In regard to this book also the
evangelical princes delivered numerous opinions, and now at
last, in obedience to the order of the princes, Andreä,
Chemnitz, Selnecker (§ 142, 4), Chytræus, Musculus, and
Körner retired into the cloister of Berg at Magdeburg in order
to make a final revision of all that was before them. Thus
originated, in A.D. 1577, the Book of Berg or the Formula of
Concord, in two different forms, first in the most compressed
style possible in what is known as the Epitome, and then
more completely in the document known as the Solida
declaratio. This document dealt with all the controverted
questions that had been agitated since A.D. 1530 in twelve
articles. It set forth the doctrine of the Person of Christ,
giving prominence to the theory of ubiquity, as the basis of
the doctrine of the supper, leaving it, however, undetermined
in accordance with the teaching of Brenz, whether the
ubiquity is to be regarded as an absolute or as a relative one,
if only it be maintained that Christ in respect of His human
nature, therefore in respect of His body, is present “ubicunque
velit,” more particularly in the holy supper. An opportunity was
also found in treating of the synergistic questions to set forth
the doctrine of predestination, although within the Lutheran
church no real controversy on this subject had ever arisen.
Luther, who at first (§ 125, 3) had himself given expression to
a particularist doctrine of election, had gradually receded
from that position. It was so too with Melanchthon, only with
this important difference, that whereas Luther, afterwards as
well as before, excluded every sort of co-operation of man in
conversion, Melanchthon felt himself obliged to admit a
certain degree of co-operation, which even the censure of
Calvin himself could not lead him to repudiate. When now the
Formula of Concord, rejecting synergism in the most decided
manner, affirmed that since the fall there was in men not
even a spark remaining, ne scintillula quidem, of spiritual
power for the independent free appropriation of offered
grace, it had gone over from the platform of Melanchthon to
that which Calvin, following the course of hard, logical
consistency, had been driven to adopt, in the assertion of a
doctrine of absolute predestination. The formula was thus in
the main in agreement with the speculation of Calvin. But it
declined to accept the conclusions arrived at in Calvinism by
declaring that while man indeed of himself wanted the power
to lay hold upon Divine grace and co-operate with it in any
way, he was yet able to withstand it and refuse to accept it.
In this way it was able to hold by the express statements of
Scripture which represent God as willing that all men should
be saved, and salvation as an absolute work of grace, but
condemnation as the consequence of man’s own guilt. It
regards the salvation of men as the only object of Divine
predestination, condemnation as merely an object of the
Divine foreknowledge.―At a later period an attempt was
made to set at rest the scruples that prevailed here and there
by securing at Berg, in February, A.D. 1580, the adoption of an
addition to it in the form of a Præfatio drawn up by Andreä as
a final determination of the controversy. The character of this
new symbolical document, in accordance with its occasion
and its aim, was not so much that of a popular exposition for
the church, but rather that of a scientific theological treatise.
For that period of excitement and controversy it is quite
remarkable and worthy of high praise for its good sense,
moderation, and circumspection, as well as for the accuracy
and clearness with which it performed its task. The fact that
nine thousand of the teachers of the church subscribed it
affords sufficient proof of it having fulfilled the end
contemplated. Denmark and Sweden, Holstein, Pomerania,
Hesse, and Anhalt, besides eight cities, Magdeburg, Dantzig,
Nuremberg, Strassburg, etc., refused to sign from various and
often conflicting motives. In A.D. 1581 Frederick II. of
Denmark is said indeed to have thrown it into the fire. Yet in
later years it was adopted in not a few of these regions, e.g.
in Sweden, Holstein, Pommerania [Pomerania], etc. The
Elector Augustus of Saxony, in the Book of Concord, brought
out a collection of all general Lutheran confessional writings
which, signed by fifty-one princes and thirty-five cities, was
solemnly promulgated on the anniversary of the Augsburg
Confession, 25th June, A.D. 1580. By this means the whole
Lutheran church of Germany obtained a common corpus
doctrinæ, and the numerous collections of confessional and
doctrinal documents acknowledged by the church, which
hitherto separate national churches had drawn up for this
purpose, henceforth lost their authority.

§ 141.13. Second Stage of Cryptocalvinism, A.D. 1586-

1592.―Yet once more the Calvinising endeavours of the
Philippists were renewed in the electorate of Saxony under
Augustus’ successor Christian I., who had obtained this
position in A.D. 1586, through his relationship with the family
of the count-palatine. His chancellor Nicholas Crell filled the
offices of pastors and teachers with men of his own views,
abolished exorcism at baptism, and had even begun the
publication of a Bible with a Calvinising commentary when
Christian died, in A.D. 1591. The Duke Frederick William of
Altenburg, as regent during the minority, immediately re-
introduced strict Lutheranism, and, preparatory to a church
visitation, had a new anti-Calvinistic standard of doctrine
compiled in the so called Articles of Visitation of A.D. 1592,
which all civil and ecclesiastical officers in Saxony were
required to accept. In short, clear, and well defined theses
and antitheses the doctrinal differences on the supper, the
Person of Christ, baptism, and election were there set forth.
In reference to baptism, the anti-Calvinistic doctrine was
promulgated, that regeneration takes place through baptism,
and that therefore every baptized person is regenerate. The
most important among the compilers of these Articles of
Visitation was Ægidius Hunnius, shortly before called to
Wittenberg, after having, from A.D. 1576 to 1592, as professor
at Marburg, laboured with all his might in opposition to the
Calvinising of Hesse. He had also, by his defence of the
doctrine of ubiquity, in his “Confession of the Doctrine of the
Person of Christ” in German, in A.D. 1577, and his Latin
treatise, “Libelli IV. de pers. Chr. ejusque ad dexteram
sedentes divina majestate,” in A.D. 1585, shown himself an
energetic champion of strict Lutheranism. He died in
A.D. 1603.―The unfortunate chancellor Crell, however, who
had made himself hateful to the Lutherans as the promoter
and chief instigator of all the Calvinising measures of the
deceased elector, and yet more so by his energetic
interference with the usurpations of the nobles, suffered an
imprisonment of ten years in the fortress of Königstein, and
was then, after a trial conducted in the most arbitrary
manner, declared to be a traitor and an enemy of the public
peace, and executed in A.D. 1601.
§ 141.14. The Huber Controversy, A.D. 1588-
1595.―Samuel Huber, reformed pastor in the Canton Bern,
became involved in a controversy with Wolfgang Musculus
over the doctrine of election. Going even beyond the Lutheran
doctrine, he affirmed that all men are predestinated to
salvation, although through their own fault not all are saved.
Banished from Bern in A.D. 1588, after a disputation with
Beza, he entered the Lutheran church and became pastor at
Württemberg. Here he charged the Professor Gerlach with
Cryptocalvinism, because he taught that only believers are
predestinated to salvation. The controversy was broken off by
his call to Wittenberg. But even his Wittenberg colleagues,
Polic. Leyser and Ægidius Hunnius, fell under the suspicion of
Cryptocalvinism, and were accordingly opposed by him. When
all disputation and conferences had failed to get him to
abandon his doctrine, and parties began to be formed among
the students, he was, in A.D. 1594, removed from Wittenberg.
With increasing rancour he continued the controversy, and
wandered about Germany for many years in order to secure a
following for his theory, but without success. He died in
A.D. 1624.

§ 141.15. The Hofmann Controversy in Helmstadt,

A.D. 1598.―The great influence which the study of the
Aristotelian philosophy in connection with that of humanism
obtained in the Julius University founded at Helmstadt in
A.D. 1576, seemed to its theological professor, Daniel
Hofmann, to threaten injury to theological study, and to be
prejudicial to pure Lutheran doctrine. He therefore attached
himself to the Romists (§ 143, 6), and took advantage of the
occasion of the conferring of doctor’s degrees to deliver a
violent invective against the incursions of reason and
philosophy into the region of religion and revelation. In
consequence of this his philosophical colleagues complained
of him to the senate as a reproacher of reason, and as one
injurious to their faculty. That court obliged him to retract and
apologise, and then deprived him of his office as professor of
theology.
 § 142. Constitution, Worship, Life, and
Science in the Lutheran Church.

In reference also to the ecclesiastical

constitution, by holding firmly to the standpoint and
to the working out of the system which it had
sketched out in its confession and doctrinal teaching,
the Lutheran church sought to mediate between
extremes, although, amid the storms from without
and from within by which it was threatened, it was
just at this point that it was least successful. It
reflected its character more clearly and decidedly in
its order of worship than in its constitution.―The
Reformation at last relaxed that hierarchical ban
which for centuries had put an absolute restraint
upon congregational singing, and had excluded the
use of the vernacular in the services of the church.
Even within the limits of the Reformation era, the
German church song attained unto such a wonderful
degree of excellence, as affords the most convincing
evidence of the fulness, power, and spirituality, the
genuine elevation and fresh enthusiasm, of the
spiritual life of that age. The sacred poetry of the
church is the confession of the Lutheran people, and
has accomplished even more than preaching for
extending and deepening the Christian life of the
evangelical church. No sooner had a sacred song of
this sort burst forth from the poet’s heart, than it was
everywhere taken up by the Christian people of the
land, and became familiar to every lip. It found
entrance into all houses and churches, was sung
before the doors, in the workshops, in the market-
places, streets, and fields, and won at a single blow
whole cities to the evangelical faith.―The Christian
life of the people in the Lutheran church combined
deep, penitential earnestness and a joyfully confident
consciousness of justification by faith with the most
nobly steadfast cheerfulness and heartiness natural
to the German citizen. Faithful attention to the
spiritual interests of their people, vigorous ethical
preaching, and zealous efforts to promote the
instruction of the young on the part of their pastors,
created among them a healthy and hearty fear of
God, without the application of any very severe
system of church discipline, a thorough and genuine
attachment to the church, strict morality in domestic
life, and loyal submission to civil
authority.―Theological science flourished especially
at the universities of Wittenberg, Tübingen,
Strassburg, Marburg, and Jena.
§ 142.1. The Ecclesiastical Constitution.―As a mean
between hierarchism and Cæsaro-papism, between the
intrusion of the State into the province of the church, and the
intrusion of the church into the province of the State, the
ecclesiastical constitution of the Lutheran church was
theoretically right in the main, though in practice and even in
theory many defects might be pointed out. It presented at
least a protest against all commingling or subordinating of
one or the other in these two spheres. Owing to the urgent
needs of the church, the princes and magistrates, in the
character of emergency-bishops, undertook the supreme
administration and management of ecclesiastical affairs, and
transferred the exercise of these rights and duties to special
boards called consistories, made up of lay and clerical
members, which were to have jurisdiction over the clergy, the
administration of discipline, and the arranging and enforcing
of the marriage laws. What had been introduced simply as a
necessity in the troubled condition of the church in those
times came gradually to be claimed as a prescriptive right.
According to the Episcopal System, the territorial lord as such
claimed to rank and act as summus episcopus. After
introducing some cautious modifications that were absolutely
indispensable, the canon law actually left the foundation of
jurisprudence untouched. The restoration of the biblical idea
of a universal priesthood of all believers would not tolerate
the retaining of the theory of an essential distinction between
the clergy and the laity. The clergy were properly designated
the servants, ministri, of the church, of the word, of the altar,
and all restrictions that had been imposed upon the clergy,
and distinguished them as an order, were removed.
Hierarchical distinctions among the clergy were renounced, as
opposed to the spirit of Christianity; but the advantage of a
superordination and subordination in respect of merely
human rights, in the institution of such offices as those of
superintendents, provosts, etc., was
recognised.―Ecclesiastical property was in many cases
diverted from the church and arbitrarily appropriated by the
greed and rapacity of princes and nobles, but still in great
part, especially in Germany, it continued in the possession of
the church, except in so far as it was applied to the
endowment of schools, universities, and charitable
institutions. The monasteries fell under a doom which by
reason of their corruptions they had richly deserved. A
restoration of such establishments in an evangelical spirit was
not to be thought of during a period of convulsion and
revolution.―Continuation, § 165, 5.

§ 142.2. Public Worship and Art.―While the Roman

Catholic order of worship was dominated almost wholly by
fancy and feeling, and that of the reformed church chiefly by
the reason, the Lutheran church sought to combine these two
features in her services. In Romish worship all appealed to
the senses, and in that of the Calvinistic churches all appealed
to the understanding; but in the Lutheran worship both sides
of human nature were fully recognised, and a proportionate
place assigned to each. The unity of the church was not
regarded as lying in the rigid uniformity of forms of worship,
but in the unity of the confession. Altars ornamented with
candles and crucifixes, as well as all the images that might be
in churches, were allowed to remain, not as objects of
worship, but rather to aid in exciting and deepening devotion.
The liturgy was closely modelled upon the Romish ritual of
the mass, with the exclusion of all unevangelical elements.
The preaching of the word was made the central point of the
whole public service. Luther’s style of preaching, the noble
and powerful popularity of which has probably never since
been equalled, certainly never surpassed, was the model and
pattern which the other Lutheran preachers set before
themselves. Among these, the most celebrated were Ant.
Corvin, Justus Jonas, George Spalatin, Bugenhagen, Jerome
Weller, John Brenz, Veit Dietrich, J. Mathesius, Martin
Chemnitz. It was laid down as absolutely essential to the idea
of public worship, that the congregation should take part in it,
and that the common language of the people should be
exclusively employed. The adoration of the sacrament on the
altar, as well as the Romish service of the mass, were set
aside as unevangelical, and the sacrament of the supper was
to be administered to the whole congregation in both kinds.
On the other hand, it was admitted that baptism was
necessary, and might and should be administered in case of
need by laymen. The customary formulary of exorcism in
baptism was at first continued without dispute, and though
Luther himself attached no great importance to it, yet every
attempt to secure its discontinuance was resisted by the later
Gnesio-Lutherans as savouring of Cryptocalvinism. Yet it
should be remembered that such orthodox representatives of
Lutheranism as Hesshus, Ægidius Hunnius, and Martin
Chemnitz, as well as afterwards John Gerhard, Quenstedt,
and Hollaz, were only in favour of its being allowed, but not
of its being regarded as necessary. Spener again declared
himself decidedly in favour of its being removed, and in the
eighteenth century it passed without any serious opposition
into disuse throughout almost the whole of the Lutheran
church, until re-introduced in the nineteenth century by the
Old Lutherans (§ 176, 2).―The church festivals were
restricted to celebrations of the facts of redemption; only
such of the feasts of Mary and the saints were retained as
had legitimate ground in the Bible history; e.g. the days of
the apostles, the annunciation of Mary, Michael’s Day,
St. John’s Day, etc. Art was held by Luther in high esteem,
especially music. Lucas Cranach, who died in A.D. 1553, Hans
Holbein, father and son, and Albert Dürer, who died in
A.D. 1528, placed their art as painters at the service of the
gospel, and adorned the churches with beautiful and
thoughtful pictures.

§ 142.3. Church Song.―The character common to the

sacred songs of the Lutheran church of the sixteenth century
is that they are thoroughly suited for congregational
purposes, and are truly popular. They are songs of faith and
the creed, with a clear impress of objectivity. The writers of
them do not describe their subjective feelings, nor their
individual experiences, but they let the church herself by their
mouths express her faith, her comfort, her thanksgiving, and
adoration. But they are also genuinely songs of the people;
true, simple, hearty, bright, and bold in expression, rapid in
movement, no standing still and looking back, no elaborate
painting and describing, no subtle demonstrating and
teaching. Even in outward form they closely resemble the old
German epics and the popular historical ballad, and were
intended above all not merely to be read, but to be sung, and
that by the whole congregation. The ecclesiastical authorities
began to introduce hymn-books into the several provinces
toward the end of the seventeenth century. Previously there
had only been private collections of sacred songs, and the
hymns were distinguished only by the words of the opening
line; and so widely known were they, that the mentioning of
them was sufficient to secure the hymn so designated being
sung by the congregation present at the public service.―The
sacred songs of the Reformation age possess all these
characteristics in remarkable degree. Among all the sacred
poets of that time Luther stands forth pre-eminent. His
thirty-six hymns or sacred poems belong to five different
classes.

1. There are free translations of Latin hymns: “Praised be

Thou, O Jesus Christ;” “Thou who art Three in unity;” “In
our true God we all believe;” “Lord God, we praise do
Thee;” “In the midst of life we are aye in death’s
embraces;” “Come God, Creator, Holy Ghost,” etc.
2. There are reproductions of original German songs:
“Death held our Lord in prison;” “Now pray we to the
Holy Ghost;” “God the Father with us be;” “Let God be
praised, blessed, and uplifted.”
3. We have also paraphrastic renderings of certain psalms:
“Ah, God in heaven, look down anew” (Ps. xii.); “Although
the mouth say of the unwise” (Ps. xiv.); “Our God, He is a
 castle strong” (Ps. xlvi.); “God, unto us right gracious be”
(Ps. lxvii.); “Had God not been with us this time”
(Ps. cxxiv.); “From trouble deep I cry to Thee” (Ps. cxxx.),
etc.
4. We have also songs composed on particular Scripture
themes: “There are the holy ten commands;” “To Isaiah
the prophet this was given” (Isa. vi.); “From heaven on
high I come to you” (Luke ii.); “To Jordan, where our
Lord has gone,” etc.
5. There are, finally, poems original in form and contents:
“Dear Christians, let us now rejoice;” “Jesus Christ, our
Saviour true;” “Lord, keep us by Thy word in hope.”405

After Luther, the most celebrated hymn-writers in the

Lutheran church of the sixteenth century are Paul Speratus,
reformer in Prussia, who died in A.D. 1554; Nicholas Decius,
first a monk, then evangelical pastor at Stettin about
A.D. 1524. Paul Eber, professor and superintendent in
Wittenberg, who died in A.D. 1569, author of the hymns,
“When in the hour of utmost need;” “Lord Jesus Christ, true
Man and God;” and one of which our well-known “Jesus, Thy
blood and righteousness,” is a paraphrase.406 Hans Sachs,
shoemaker in Nuremberg, who died in A.D. 1567, wrote during
the famine in that city in A.D. 1552 the hymn, “Why art thou
thus cast down, my heart?” John Schneesing, pastor in
Gothaschen, who died in A.D. 1567, wrote “Lord Jesus Christ,
in Thee alone.” John Mathesius, rector and deacon in
Joachimsthal, who also delivered sermons on Luther’s life,
died in A.D. 1565, wrote a beautiful morning hymn, and other
sweet sacred pieces. Nicholas Hermann, who died in
A.D. 1561, precentor at Joachimsthal, wrote out Mathesius’
sermons in hymns, “The happy sunshine all is gone,” the
burial hymn, “Now hush your cries, and shed no tear,” etc.
Michael Weisse closes the series of hymn-writers of the
Reformation age. He was a German pastor in Bohemia,
translator and editor of the sacred songs of the Bohemian
Hussites, and died in A.D. 1540. He wrote “Christ the Lord is
risen again,” and the burial hymn to which Luther added a
verse, “Now lay we calmly in the grave.”407

§ 142.4. In the period immediately following, from A.D. 1560

to A.D. 1618, we meet with many poetasters who write on
sacred themes in doggerel rhymes. Even those who are poets
by natural endowment, and inspired with Divine grace, are
much too prolific; but they have bequeathed to us a genuine
wealth of beautiful church songs, characterized by healthful
objectivity, childlike simplicity, and a singular power of
appealing to the hearts of the great masses of the people.
But a tendency already begins to manifest itself in the
direction of that excessive subjectivity which was the vice of
hymn-writers in the succeeding period; the doctrinal element
too becomes more and more prominent, as well as
application to particular circumstances and occasions in life;
but the objective confession of faith is always still
predominant. Among the sacred poets of this period the most
important are Bartholmaus Ringwaldt, pastor in
Brandenburg, who died in A.D. 1597, author of “’Tis sure that
awful time will come;” Nicholas Selnecker, at last
superintendent in Leipzig, who died in A.D. 1592, as
Melanchthon’s scholar suspected at one time of
Cryptocalvinism, but, after he had taken part in the
composition of the Formula of Concord, the object of the
most bitter hatred and constant persecution on the part of
the Cryptocalvinists of Saxony: he wrote, “O Lord my God, I
cry to Thee;” Martin Schalling, pastor at Regensburg and
Nuremberg, who died in A.D. 1608, wrote, “Lord, all my heart
is fixed on Thee;” Martin Böhme or Behemb, pastor in
Lusatia, who died in A.D. 1621, author of “Lord Jesus Christ,
my Life, my Light.” The series closes with Philip Nicolai, a
violent and determined opponent of Calvinism, who was
latterly pastor in Hamburg, and died in A.D. 1608. His vigorous
and rhythmical poetry, with its deep undertone of sweetness,
is to some extent modelled on the Song of Songs. He wrote
“Awake, awake, for night is flying;” the chorale in
Mendelssohn’s “St. Paul,” “Sleepers, wake, a voice is calling,”
is a rendering of the same piece.―Continuation, § 159, 3.

§ 142.5. Chorale Singing.―The congregational singing,

which the Reformation made an integral part of evangelical
worship, was essentially a reproduction of the Ambrosian
mode (§ 59, 5) in a purer form and with richer fulness. It was
distinguished from the Gregorian style preeminently by this,
that it was not the singing of a choir of priests, but the
popular singing of the whole congregation. The name chorale
singing, however, was still continued, and has come to be the
technical and appropriate designation of the new mode. It is
further distinguished from the Gregorian mode by this other
characteristic, that instead of singing in a uniform monotone
of simple notes of equal length, it introduces a richer rhythm
with more lively modulation. And, finally, it is characterized by
the introduction of harmony in place of the customary unison.
But, on the other hand, the chorale singing may be regarded
as a renewal of the old cantus firmus, while at the same time
it sets aside the secular music style and the artificialities of
counterpoint and the elaborate ornamentation with which the
false taste of the Middle Ages had overlaid it. The
congregation sang the cantus firmus or melody in unison, the
singers in the choir gave it the accompaniment of a harmony.
The organ during the Reformation age was used for support,
and accompanied only in elaborate, high-class music. But the
melody was pitched in a medium key, which as the leading
voice was called Tenor. The melodies for the new church
hymns were obtained, partly by adaptation of the old tunes
for the Latin hymns and sequences, partly by appropriation of
popular mediæval airs, especially among the Bohemian
Brethren, partly also and mainly by the free use of the
popular song tunes of the day, to which no one made any
objection, since indeed the spiritual songs were often
parodies of the popular songs whose airs were laid hold upon
for church use. The few original melodies of this age were for
the most part composed by the authors of the hymns
themselves or by the singers, and were the outflow of the
same inspiration as had called forth the poems. They have
therefore been rarely equalled in impressiveness, spiritual
glow, and power by any of the more artistic productions of
later times. Acquaintance with the new melodies was spread
among the people by itinerant singers, chorister boys in the
streets, and the city cornet players. From the singers or those
who adapted the melodies are to be distinguished the
composers, who as technical musicians arranged the harmony
and set it in a form suitable for church use. George Rhaw,
precentor in Leipzig, afterwards printer in Wittenberg, and
Hans Walter, choirmaster to the elector, both intimate
friends of Luther, were amongst the most celebrated
composers of their day. The evangelical church music reaches
its highest point of excellence toward the end of the sixteenth
century. The great musical composer, John Eccart, who was
latterly choirmaster in Berlin, and died in A.D. 1611, was the
most active agent in securing this perfection of his art. In
order to make the melody clearer and more distinctly heard, it
was transferred from the middle voice, the tenor, to the
higher voice or treble. The other voices now came in as
simple concords alongside of the melody, and the organ,
which had now been almost perfected by the introduction of
many important improvements, now came into general use
with its pure, rich, and accurate full harmony, as a support
and accompaniment of the congregational singing. The
distinction too between singers and composers passed more
and more out of view. The skilled artistic singing was thus
brought into closer relations with the congregational singing,
and the creative power, out of which an abundant supply of
original melodies was produced, grew and developed from
year to year.
§ 142.6. Theological Science.―Inasmuch as the
Reformation had its origin in the word of God, and supported
itself upon that foundation alone the theologians of the
Reformation were obliged to give special attention to biblical
studies. John Förster, who died in A.D. 1556, and John
Avenarius, who died in A.D. 1576, both of Wittenberg,
compiled Hebrew lexicons, which embodied the results of
independent investigations. Matthias Flacius, in his Clavis
Scr. s., provided what for that time was a very serviceable aid
to the study of Scripture. The first part gives in alphabetical
order an explanation of Scripture words and forms of speech,
the second forms a system of biblical hermeneutics. Exegesis
proper found numerous representatives. Luther himself
beyond dispute holds the front rank in this department. After
him the most important Lutheran exegetes of that age are for
the New Testament, Melanchthon; Victorin Strigel, who wrote
Hyponm. in Novum Testamentum; Flacius, with his Glossa
compendiaria in Novum Testamentum; Joachim Camerarius,
with his Notationes in Nov. Testamentum; Martin Chemnitz,
with his Harmonia IV. Evangeliorum, continued by Polic.
Leyser, and completed at last by John Gerhard: for the Old
Testament, especially John Brenz, whose commentaries are
still worthy of being consulted. Of less consequence are the
numerous commentaries of the comprehensive order,
compiled by the once scarcely less influential David Chytræus
of Rostock, who died in A.D. 1600. The series of Lutheran
dogmatists opens with Melanchthon, who published his Loci
communes in A.D. 1521. Martin Chemnitz, in his Loci
theologici, contributed an admirable commentary to
Melanchthon’s work, and it soon became the recognised
standard dogmatic treatise in the Lutheran church. In
A.D. 1562 he published his Examen Conc. Trident., in which he
combated the Romish doctrine with as much learning and
thoroughness as good sense, mildness, and moderation.
Polemical theology was engaged upon with great vigour amid
the many internal and external controversies, conducted often
with intense passion and bitterness. In the department of
church history we have the gigantic work of the Magdeburg
centuriators, the result of the bold scheme of Matthias
Flacius. By his Catalogus testium veritatis he had previously
advanced evidence to show that at no point in her history had
the church been without enlightened and pious heroes of
faith, who had carried on the uninterrupted historical
continuity of evangelical truth, and so secured an unbroken
succession from the early apostolic church till that of the
sixteenth century.―Continuation, § 159, 4.

§ 142.7. German National Literature.―The Reformation

occurred at a time when the poetry and national literature of
Germany was in a condition of profound prostration, if not
utter collapse. But it brought with it a reawakening of creative
powers in the national and intellectual life of the people.
Under the influence and stimulus of Luther’s own example
there arose a new prose literature, inspired by a broad, liberal
spirit, as the expression of a new view of the world, which led
the Germans both to think and teach in German. It was
mainly the intellectual friction from the contact of one fresh
mind with another in regard to questions agitated in the
Reformation movement that gave to the satirical writings of
the age that brilliancy, point, and popularity which in the
history of German literature was not attained before and
never has been reached since. In innumerable fugitive sheets,
in the most diverse forms of style and language, in poetry
and prose, in Latin and German, these satires poured forth
contempt and scorn against and in favour of the Reformation.
As we have on the Catholic side Thomas Murner (§ 125, 4),
and on the Reformed side Nicholas Manuel (§ 130, 4), so we
have on the Lutheran side John Fischart, far excelling the
former two, and indeed the greatest satirist that Germany has
yet produced. To him we are mainly indebted for the almost
incessant stream of anonymous satires of the sixteenth
century. He belonged, like Sebastian Brandt and Thomas
Welcome to Our Bookstore - The Ultimate Destination for Book Lovers
Are you passionate about books and eager to explore new worlds of
knowledge? At our website, we offer a vast collection of books that
cater to every interest and age group. From classic literature to
specialized publications, self-help books, and children’s stories, we
have it all! Each book is a gateway to new adventures, helping you
expand your knowledge and nourish your soul
Experience Convenient and Enjoyable Book Shopping Our website is more
than just an online bookstore—it’s a bridge connecting readers to the
timeless values of culture and wisdom. With a sleek and user-friendly
interface and a smart search system, you can find your favorite books
quickly and easily. Enjoy special promotions, fast home delivery, and
a seamless shopping experience that saves you time and enhances your
love for reading.
Let us accompany you on the journey of exploring knowledge and
personal growth!

ebookball.com