Guide to Computational Modelling for Decision Processes. Theory, Algorithms, Techniques and Applications

384 Pages·2017·8.207 MB·english

by  Stuart Berry, Val Lowndes, Marcello Trovati

Stuart Berry Val Lowndes
(cid:129)
Marcello Trovati
Editors
Guide to
Computational Modelling
for Decision Processes
Theory, Algorithms,
Techniques and Applications
123
Editors
Stuart Berry Marcello Trovati
Department ofComputing andMathematics University of Derby
Collegeof Engineering andTechnology Derby
University of Derby UK
Derby
UK
ValLowndes
University of Derby
Derby
UK
ISSN 2195-2817 ISSN 2195-2825 (electronic)
Simulation Foundations,Methods andApplications
ISBN978-3-319-55416-7 ISBN978-3-319-55417-4 (eBook)
DOI 10.1007/978-3-319-55417-4
LibraryofCongressControlNumber:2017934055
©SpringerInternationalPublishingAG2017
ThisSpringerimprintispublishedbySpringerNature
TheregisteredcompanyisSpringerInternationalPublishingAG
Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland
Preface
Outline of Content
This book is organised into three sections, Part I introduces modelling techniques
and models used to represent application used later to evaluate solution processes
and procedures.
Part I: Introduction to Modelling and Model Evaluation.
Thispartintroducesmodellingmethodologiesandmodelstobeusedasstarting
points to enable the derivation of efficient and effective solution techniques.
Part II: Case Studies.
This part presents a series of case studies to demonstrate how heuristic and
analytical approaches may be used to solve large complex problems.
A series of case studies are presented where models are constructed and then
analysed and evaluated to derive efficient and effective ways to produce good
solutions.
Within Part I:
Chapter 1: Model Building. This chapter introduces the modelling
methodologies:
Activity life cycles and problem analysis using activity life cycles.
Constructing models from “Big Data”.
Blackboard modelling.
Chapter 2: Introduction to Cellular Automata in Simulation. This chapter aims
to show how both these approaches can be used as modelling tools.
Introduction to cellular automata.
Cellular automata are introduced by way of Conway’s Game of Life and
applying agents.
Simulating the spread of an infection through a population.
Traffic modelling, investigating traffic flows.
Chapter 3: Introduction to Mathematical Programming. The aim in this chapter
istodemonstratehowamathematicalprogrammingmodelcanbeusedtodescribe
and explain the complexity of a planning problem and hence lead towards an
efficient solution technique or methodology.
Whether (these) problems have easy solutions or because of the inherent com-
plexity of the form of the required solution, they are better approached using
heuristic techniques (generally accepting good rather than best solutions).
To achieve this objective, the following problems are presented:
Diet problems, the very obvious formulations (Stigler and Dantzig), lead to an
undesirable solution (only one meal!), and a more heuristic approach is needed
to add incorporate multi-objectives that typically minimise cost while max-
imising variety/taste.
Knapsack problems, showing how many problems are reducible to knapsack
problems and are therefore appropriate for the use of heuristic solution
techniques.
Network flow problems, again many planning problems can be modelled as
network flow problems and hence can be solved easily.
Chapter 4: Heuristic Techniques in Optimisation. This part introduces approa-
chesthatcanbeusedtoobtaingoodsolutionstohardorlargeproblemscomparing
and contrasting the effectiveness and efficiency of heuristic approaches to
problem-solving.
To achieve this objective, the following approaches are presented:
Genetic algorithms implementations illustrated through its application in pro-
ducing solutions to knapsack problems, travelling salesman problems,
scheduling problems, and quadratic assignment problems.
Tabu search implementations illustrated through its application in producing
solutions to financial planning and travelling salesman problems.
Chapter 5: Introduction to the Use of Queueing Theory and Simulation. This
chapter shows how queueing theory and simulation techniques can be applied to
design efficient and effective service systems.
The major sections are concerned with the following:
Queueing theory leading to “quick” modelling, how queueing theory can be
employed to carry out an evaluation of a manufacturing system.
Simulation modelling, introducing an alternative approach to modelling com-
plex planning problems.
Part II: Case Studies
This part presents a series of case studies to demonstrate how heuristic and
analytical approaches may be used to solve large complex problems.
A series of case studies are presented where models are constructed and then
analysed and evaluated to derive efficient and effective ways to produce good
solutions.
Maybesolvedtypicalsuggestedapplications,presentingalternativeapproaches
to problem solving
Chapter 6 describes an investigation into the appropriateness of heuristic
methodologies in the solution of
Travelling salesman problem.
Garbage collection problem, a multiple travelling salesman (type) problem.
Productionplanningandcontrolproblems,whereaseeminglyhardproblemcan
beshowntobesolvableusingasimpleheuristicapproachreducingtheneedfor
cost data.
Chapter 7 describes how an efficient heuristic approach can be derived from an
initial complex model using the following:
Flow shop scheduling, showing how a hard problem can be approached using
heuristic methods.
Transport planning, deriving approaches to evaluate the benefits to be gained
from the installation of an active traffic control system and the paradoxes
resulting from changes to transport planning.
Chapter 8 describes an investigation into the production of an efficient and
effective means of scheduling air traffic controllers, where the method used has to
have the ability to respond (create a new schedule) rapidly to staff availabilities.
Chapter 9 describes how a multiple objective optimisation problem can be
solved by the incorporation of techniques from genetic algorithms and fuzzy logic
into a mathematical programming methodology.
Thisapproachisillustratedbyitsapplicationintotheprovisionofsolutionstoa
diet problem with the extended objectives:
Minimise cost.
Produce a healthy diet.
Produce a large variety of good diets.
Chapter10describesaninvestigationintotheapplicationoffuzzylogicshowing
how it can be used to derive a dynamic method of scheduling operations in a
workshop.
This approach is applied to a workshop where there are multiple objectives:
Importanceofcustomeranddeliveryduedates,usingfuzzylogictoderivework
schedules.
Chapter11describes howanapproachbasedontabusearch methodologiescan
be employed to derive optimal control settings.
This approach is illustrated through its application to a Surround Sound 5
speaker system determining settings so that the system could produce “perfect”
directional sound.
Chapter 12 describes how system dynamics modelling can be employed to
describe the output from complex decision making processes.
Models are constructed to describe
the changes in the Dow Jones index, from growth to decline, and
thechangesinthedominantmodeoftransport(withtime)andtheeffectofthese
changes on the prior dominant modes.
Chapter 13 describes the use of queueing theory in the evaluation of traffic
controlsystems(trafficlights)showinghowthesystemcouldbeimprovedthrough
the use of “available forward road capacity” that is passing information between
traffic lights.
Chapter 14 describes case study investigations into the use of cellular automata
and agent-based simulations.
The case studies are based on message passing, by mobile devices, within a
closed environment (a shopping centre for example) and
The spread of a fire and the improved positioning of the fire exits in a closed
environment.
Chapter 15 discusses the use of “Big Data” to derive models.
Three case studies are provided.
Criminology,
Depression evaluation, and
University admissions.
Derby, UK Stuart Berry
Val Lowndes
Marcello Trovati
Contents
Part I Introduction to Modelling and Model Evaluation
1 Model Building .... .... ..... .... .... .... .... .... ..... .... 3
ValLowndes,StuartBerry,MarcelloTrovatiandAmandaWhitbrook
2 Introduction to Cellular Automata in Simulation.. .... ..... .... 55
Val Lowndes, Adrian Bird and Stuart Berry
3 Introduction to Mathematical Programming.. .... .... ..... .... 75
Val Lowndes and Stuart Berry
4 Heuristic Techniques in Optimisation ... .... .... .... ..... .... 121
Val Lowndes and Stuart Berry
5 Introduction to the Use of Queueing Theory and Simulation.. .... 145
Val Lowndes and Stuart Berry
Part II Case Studies
6 Case Studies: Using Heuristics. .... .... .... .... .... ..... .... 175
Val Lowndes, Ovidiu Bagdasar and Stuart Berry
7 Further Use of Heuristic Methods .. .... .... .... .... ..... .... 199
Val Lowndes, Stuart Berry, Chris Parkes, Ovidiu Bagdasar
and Nicolae Popovici
8 Air Traffic Controllers Planning: A Rostering Problem. ..... .... 237
Richard Conniss
9 Solving Multiple Objective Problems: Modelling
Diet Problems . .... .... ..... .... .... .... .... .... ..... .... 251
Val Lowndes and Stuart Berry
10 Fuzzy Scheduling Applied to Small Manufacturing Firms.... .... 265
Val Lowndes
11 The Design and Optimisation of Surround Sound Decoders
Using Heuristic Methods. ..... .... .... .... .... .... ..... .... 273
Bruce Wiggins, Stuart Berry and Val Lowndes
12 System Dynamics Case Studies. .... .... .... .... .... ..... .... 285
Chris Parkes, Stuart Berry and John Stubbs
13 Applying Queueing Theory to the Design of a Traffic Light
Controller. .... .... .... ..... .... .... .... .... .... ..... .... 299
James Hardy
14 Cellular Automata and Agents in Simulations .... .... ..... .... 307
Kim Smith, Richard Hill, Stuart Berry and Richard Conniss
15 Three Big Data Case Studies .. .... .... .... .... .... ..... .... 333
Marcello Trovati and Andy Baker
Part III Appendices
16 Appendix A: Queueing Theory. .... .... .... .... .... ..... .... 349
Stuart Berry
17 Appendix B: Function Optimisation Techniques Genetic
Algorithms and Tabu Searches .... .... .... .... .... ..... .... 355
Val Lowndes and Mirko Paskota
18 Appendix C: What to Simulate to Evaluate Production Planning
and Control Methods in Small Manufacturing Firm’s.. ..... .... 377
Val Lowndes and Stuart Berry
19 Appendix D: Defining Boolean and Fuzzy Logic Operators... .... 381
Val Lowndes
20 Appendix E: Assessing the Reinstated Waverly Line... ..... .... 383
Stuart Berry and John Stubbs
21 Appendix F: Matching Services with Users in Opportunistic
Network Environments.. ..... .... .... .... .... .... ..... .... 387
Stuart Berry
References.... .... .... .... ..... .... .... .... .... .... ..... .... 391
Index .... .... .... .... .... ..... .... .... .... .... .... ..... .... 395
Part I
Introduction to Modelling
and Model Evaluation
This section introduces modelling techniques and constructs models to represent
and analyse planning problems in business, industry and the management of
facilities.
Theseconstructedmodelsareevaluated;cantheybesolvedinareasonabletime
using standard analytical techniques or should the solution be approached using
heuristic methods or heuristic methodologies?
Chapter 1
Model Building
Val Lowndes, Stuart Berry, Marcello Trovati
and Amanda Whitbrook
Section 1.1 introduces theuseofsystemdynamics inmodellingandthen uses this
approach to construct models to describe real applications.
Section 1.2 introduces the concepts needed to construct models using available
data, modelling using Big Data.
Section 1.3 introduces modelling using blackboard architecture; this provides a
flexible, symbolic artificial intelligence (AI) method for the cooperative modelling
and then solution of complex problems.
1.1 Introduction to System Modelling
The purpose of system dynamics modelling is to develop understanding and then
the improvement of systems. The first stage in this process is the construction of a
logical model (influence diagram) to describe a system.
V.Lowndes(Retired)
UniversityofDerby,KedlestonRoad,DE221GBDerby,UK
e-mail:[email protected]
S.Berry(&)(cid:1)A.Whitbrook
CollegeofEngineeringandTechnology,UniversityofDerby,
KedlestonRoad,DE221GBDerby,UK
e-mail:[email protected]
A.Whitbrook
e-mail:[email protected]
M.Trovati
ComputerScience,EdgeHillUniversity,StHelensRoad,Ormskirk,
L394QPLancashire,UK
e-mail:[email protected]