Evolutionary and Adaptive Computing in Engineering Design: With 98 Figures

Front Cover
Springer Science & Business Media, 2001 - Computers - 286 pages
Prior to the early 1990s the term 'evolutionary computing' (EC) would have meant little to most practising engineers unless they had a particular interest in emerging computing technologies or were part of an organisation with significant in-house research activities. It was around this time that the first tentative utilisation of relatively simple evolutionary algorithms within engineering design began to emerge in the UK The potential was rapidly recognised especially within the aerospace sector with both Rolls Royce and British Aerospace taking a serious interest while in the USA General Electric had already developed a suite of optimisation software which included evolutionary and adaptiv,e search algorithms. Considering that the technologies were already twenty-plus years old at this point the long gestation period is perhaps indicative of the problems associated with their real-world implementation. Engineering application was evident as early as the mid-sixties when the founders of the various techniques achieved some success with computing resources that had difficulty coping with the population-based search characteristics of the evolutionary algorithms. Unlike more conventional, deterministic optimisation procedures, evolutionary algorithms search from a population of possible solutions which evolve over many generations. This largely stochastic process demands serious computing capability especially where objective functions involve complex iterative mathematical procedures.
 

What people are saying - Write a review

We haven't found any reviews in the usual places.

Contents

Introduction
1
12 Why EvolutionaryAdaptive Computing?
2
13 The UK EPSRC Engineering Design Centres
4
14 Evolutionary and Adaptive Computing Integration
6
15 Generic Design Issues
12
16 Moving On
15
Established Evolutionary Search Algorithms
17
22 A Brief History of Evolutionary Search Techniques
18
82 Improving Rolls Royce Cooling Hole Geometry Models
134
83 Discussion of Initial Application
142
84 Further Development of the GP Paradigm
143
85 Symbolic Regression with HDRAMGP
145
86 Dualagent Integration
146
87 Return to Engineering Applications
147
88 Discussion
149
Evolutionary Constraint Satisfaction and Constrained Optimisation
151

23 The Genetic Algorithm
19
24 GA Variants
29
25 Evolution Strategies
36
26 Evolutionary Programming
38
27 Genetic Programming
39
28 Discussion
43
Adaptive Search and Optimisation Algorithms
45
32 The Antcolony Metaphor
46
33 Populationbased Incremental Learning
49
34 Simulated Annealing
51
35 Tabu Search
53
37 Discussion
56
Initial Application
59
43 The Design of Gas Turbine Blade Cooling Hole Geometries
64
44 Evolutionary FIR Digital Filter Design
71
45 Evolutionary Design of a Threecentred Concrete Arch Dam
75
46 Discussion
77
The Development of Evolutionary and Adaptive Search Strategies for Engineering Design
79
52 Clusteroriented Genetic Algorithms
80
54 DRAM and HDRAM Genetic Programming Variants
81
55 Evolutionary and Adaptive Search Strategies for Constrained Problems
83
56 Evolutionary Multicriterion Satisfaction
84
57 Designer Interaction within an Evolutionary Design Environment
85
58 Dynamic Shape Refinement and Injection Island Variants
86
59 Discussion
87
Evolutionary Design Space Decomposition
89
62 Multimodal Optimisation
90
63 Clusteroriented Genetic Algorithms
91
64 Application of vmCOGA
95
65 Alternative COGA Structures
102
66 Agentassisted Boundary Identification
107
67 Discussion
108
Wholesystem Design
111
72 Previous Related Work
114
73 The Hydropower System
115
74 The Structured Genetic Algorithm
118
75 Simplifying the Parameter Representation
121
76 Results and Discussion
124
77 Thermal Power System Redesign
126
78 Discussion
130
Variablelength Hierarchies and System Identification
133
92 Dealing with Explicit Constraints
152
93 Implicit Constraints
158
94 Defining Feasible Space
160
95 Satisfying Constraint in the Optimisation of Thermal Power Plant Design
166
96 GAAntcolony Hybrid for the Flight Trajectory Problem
168
97 Other Techniques
173
98 Discussion
174
Multiobjective Satisfaction and Optimisation
177
102 Established Multiobjective Optimisation Techniques
178
103 Interactive Approaches to Multiobjective SatisfactionOptimisation
185
104 Qualitative Evaluation of GAgenerated Design Solutions
186
105 Clusteroriented Genetic Algorithms for Multi objective Satisfaction
194
106 Related Work and Further Reading
200
107 Discussion
202
Towards Interactive Evolutionary Design Systems
205
112 System Requirements
206
113 The Design Environment and the IEDS
208
114 The Rulebased Preference Component
210
115 The Coevolutionary Environment
213
116 Combining Preferences with the Coevolutionary Approach
218
117 Clusteroriented Genetic Algorithms as Information Gathering Processes
220
118 Machinebased Agent Support
221
119 Machinebased Design Space Modification
225
1110 Discussion
230
Populationbased Search Shape Optimisation and Computational Expense
233
122 Parallel Distributed and Coevolutionary Strategies
234
123 Introducing the Problem and the Developed Strategies
237
124 The Evaluation Model
238
125 Initial Results
239
127 The Injection Island GA
241
128 Dynamic Injection
244
129 Distributed Search Techniques
246
1210 Discussion
250
Closing Discussion
253
133 Overview of the Techniques and Strategies Introduced
255
134 Final Remarks
259
Some Basic Concepts
261
References
265
Index
279
Copyright

Other editions - View all

Common terms and phrases

References to this book

Bibliographic information