Creative Evolutionary Systems
1st Edition
Table of Contents
About the Editors
Foreword
By Margaret Boden
Contributors
Preface
An Introduction to Creative Evolutionary Systems
By Peter J. Bentley and David W. Corne
Introduction
AI and Creativity
Evolutionary Computation
Creative Evolutionary Systems
Is Evolution Creative?
PART I - Evolutionary Creativity
Chapter 1 - Creativity in Evolution: Individuals, Interactions, and Environments
By Tim Taylor
1.1 Introduction
1.2 Creativity and Opened-Ended Evolution
1.3 Design Issues
1.3.1 Von Neumann's Architecture for Self-Reproduction
1.3.2 Tierra
1.3.3 Implicit versus Explicit Encoding
1.3.4 Ability to Perform Other Tasks
1.3.5 Embeddedness in the Arena of Competition and Richness of Interactions
1.3.6 Materiality
1.4 A Full Specification For An Open-Ended Evolutionary Process
1.4.1 Waddington's Paradigm for an Evolutionary Process
1.5 Conclusions
Acknowledgments
References
Chapter 2 - Recognizability of the Idea: The Evolutionary Process of Argenia
By Celestino Soddu
2.1 Introduction
2.2 Recognizability, Identity, And Complexity
2.3 Evolutionary Codes: Artificial DNA
2.4 Natural/Artificial Complexity
2.5 Giotto, A Medieval Idea In Evolution
2.6 Rome, Future Scenarios
2.7 Basilica, Generative Software To Design Complexity
2.8 Madrid and Milan, Generated Architecture
2.9 Argenìa, The Natural Industrial Object, And The Artificial Uniqueness Of Species
2.10 Argenìc Art: Picasso
2.11 Conclusions
References
Chapter 3 - Breeding Aesthetic Objects: Art and Artificial Evolution
By Mitchell Whitelaw
3.1 Introduction
3.2 Breeding Aesthetic Objects
3.2.1 A Case Study—Steven Rooke
3.3 Breeding and Creation
3.3.1 Creative Agency and the Breeding Process
3.3.2 The Evolved Aesthetic Object
3.4 Limits
3.5 Driessens and Verstappen—An Alternative Approach
3.6 Conclusions
References
Chapter 4 - The Beer Can Theory of Creativity
By Liane Gabora
4.1 Introduction
4.2 Culture As An Evolutionary Process
4.2.1 Variation and Convergence in Biology and Culture
4.2.2 Is More Than One Mind Necessary for Ideas to Evolve?
4.2.3 Meme and Variations: A Computer Model of Cultural Evolution
4.2.4 Breadth-First versus Depth-First Exploration
4.2.5 Dampening Arbitrary Associations and Forging Meaningful Ones
4.3 Creativity as The Origin Of Culture
4.3.1 Theoretical Evidence
4.3.2 Archeological Evidence
4.3.3 Evidence from Animal Behavior
4.4 What Caused the Onset of Creativity?
4.5 Conclusions
Acknowledgments
References
PART II Evolutionary Music
Chapter 5 - GenJam: Evolution of a Jazz Improviser
By John A. Biles
5.1 Introduction
5.2 Overview and Architecture
5.3 Representations
5.4 Genetic Operators and Training
5.4.1 Crossover
5.4.2 Musically Meaningful Mutation
5.5 Real-Time Interaction
5.6 Conclusions
References
Chapter 6 - On the Origins and Evolution of Music in Virtual Worlds
By Eduardo Reck Miranda
6.1 Introduction
6.2 Evolutionary Modeling
6.2.1 Transformation and Selection
6.2.2 Coevolution
6.2.3 Self-organization
6.2.4 Level Formation
6.3 Evolving Sound With Cellular Automata
6.3.1 The Basics of Cellular Automata
6.3.2 The Cellular Automaton Used in Our System
6.3.3 The Synthesis Engine
6.4 Commentary On The Results
6.5 Conclusions
Acknowledgments
References
Chapter 7 - Vox Populi: Evolutionary Computation for Music Evolution
By Artemis Moroni, Jônatas Manzolli, Fernando Von Zuben, and Ricardo Gudwin
7.1 Introduction
7.2 Sound Attributes
7.3 Evolutionary Musical Cycle
7.3.1 The Voices Population
7.3.2 The Rhythm of the Evolution
7.4 Fitness Evaluation
7.4.1 The Consonance Criterion
7.4.2 Melodic Fitness
7.4.3 Harmonic Fitness
7.4.4 Voice Range Criterion
7.4.5 Musical Fitness
7.5 Interface And Parameter Control
7.6 Experiments
7.7 Conclusions
Acknowledgments
References
Chapter 8 - The Sound Gallery—An Interactive A-Life Artwork
By Sam Woolf and Adrian Thompson
8.1 Introduction
8.2 Evolvable Hardware
8.2.1 Reconfigurable Chips
8.3 Gallery Setup
8.3.1 Setting
8.3.2 Sensing Systems
8.4 Contextualization: Artificial Life and Art
8.4.1 Evolutionary Algorithms and Visual Arts
8.4.2 Evolutionary Algorithms and Music
8.4.3 Interactive Genetic Art
8.4.4 Interactive, Adaptive, and Autonomous (Nongenetic) Artworks
8.5 The Sound Gallery Algorithms
8.5.1 Two-Phase Hill-Climbing/ Island Model GA
8.5.2 Hill-climbing Phase
8.5.3 Island Model Genetic Algorithm Phase
8.5.4 The Need for Aging
8.5.5 Encoding Scheme
8.5.6 The Fitness Function
8.5.7 galSim
8.6 The Experiment
8.6.1 Results
8.7 Conclusions
Acknowledgments
References
Contents
PART III Creative Evolutionary Design
Chapter 9 - Creative Design and the Generative Evolutionary Paradigm
By John Frazer
9.1 Introduction
9.2 The Adaptive Model From Nature
9.3 The Generative Evolutionary Paradigm
9.4 Problems With The Paradigm
9.5 Concept Seeding Approach
9.6 The Reptile Demonstration
9.7 Universal State Space Modeler
9.8 Logic Fields
9.9 Returning to the Analogy with Nature
9.10 Conclusions
References
Chapter 10 - Genetic Programming: Biologically Inspired Computation That Exhibits Creativity in Producing Human-Competitive Results
By John R. Koza, Forrest H. Bennett III, David Andre, and Martin A. Keane
10.1 Introduction
10.2 Inventiveness And Creativity
10.3 Genetic Programming
10.4 Applying Genetic Programming To Circuit Synthesis
10.4.1 Campbell 1917 Ladder Filter Patent
10.4.2 Zobel 1925 "M-Derived Half Section" Patent
10.4.3 Cauer 1934-1936 Elliptic Filter Patents
10.4.4 Amplifier, Computational, Temperature-Sensing, Voltage Reference, and Other Circuits
10.5 Topology, Sizing, Placement, and Routing Of Circuits Contents
10.6 Automatic Synthesis Of Controllers By Means Of Genetic Programming
10.6.1 Robust Controller for a Two-Lag Plant
10.7 The Illogical Nature Of Creativity And Evolution
10.8 Conclusions
References
Chapeter 11 - Toward a Symbiotic Coevolutionary Approach to Architecture
By Helen Jackson
11.1 Introduction
11.2 Lindenmayer Systems
11.2.1 Example L-Systems
11.2.2 The Isospatial Grid
11.2.3 Spatial Embryology
11.3 Artificial Selection
11.3.1 The Eyeball Test
11.4 Single-Goal Evolution
11.4.1 "Generic Function" as Fitness Function
11.4.2 Evolution toward Low i-Values
11.4.3 Structural Stability
11.4.4 Architecture As a Multigoal Task
11.4.5 Dual-Goal Evolution
11.5 Representation, Systems, And Symbiosis
11.5.1 Coevolution
11.5.2 Naïve Architectural Form Representation
11.5.3 Spatial Embryology
11.6 Conclusions
Acknowledgments
References
Chapter 12 - Using Evolutionary Algorithms to Aid Designers of Architectural Structures
By Peter von Buelow
12.1 Introduction
12.2 Analysis Tools Vs. Design Tools
12.3 Advantages Of Evolutionary Systems In Design Contents
12.3.1 Use of Populations
12.3.2 Recombination and Mutation
12.3.3 Wide Search of Design Space
12.3.4 No Knowledge of the Objective Function
12.3.5 Imitation of Human Design Process
12.3.6 Can Learn from Designer
12.4 Characteristics of an IGDT
12.4.1 Definition of the IGDT Concept
12.4.2 Relation of IGDT to Design Process
12.5 Mechanics of an IGDT
12.6 IGDT Operation
12.6.1 Problem Definition
12.6.2 Initial IGDT Generation
12.6.3 Initial Generation with Designer Selection/Interaction
12.6.4 Second-Generation IGDT Response
12.6.5 Second-Generation Designer Interaction
12.6.6 Third Generation
12.7 Conclusions
Acknowledgments
References
PART IV Evolutionary Art
Chapter 13 - Eons of Genetically Evolved Algorithmic Images
By Steven Rooke
13.1 Introduction
13.2 Using GP for Art
13.2.1 Genetic Variation
13.2.2 Genetic Library
13.2.3 Functions and Node Internals
13.2.4 A Typical Run
13.3 Horizon Lines And Fantasy Landscapes
13.4 Genetic Fractals
13.4.1 Second-Order Subtleties of Orbit Trajectories during Iteration in the Complex Plane
13.5 The Genetic Cross Dissolve
13.6 What Is It?
13.6.1 Constraints of Color and Form
13.6.2 A Joyride for the Visual Cortex?
13.6.3 Approaching the Organic
13.7 Conclusions
References
Chapter 14 - Art, Robots, and Evolution as a Tool for Creativity
By Luigi Pagliarini and Henrik Hautop Lund
14.1 Introduction
14.2 The Social Context Of Electronics
14.2.1 Where Electronics Acts
14.2.2 How Technology Influences Art (the World)
14.2.3 How Technology Gets Feedback (from Art and the World)
14.3 What Artist?
14.3.1 Two Different Concepts or Aspects of the Artist
14.3.2 Art and Human Language: The "Immaterial" Artist
14.3.3 Art and Human Technique: The "Material" Artist
14.4 Electronic Art
14.4.1 A New Electronic Space
14.4.2 The "Material" Electronic Artist
14.4.3 The "Immaterial" Artist and the Uses of Electronics
14.4.4 Example—The Artificial Painter
14.5 Alive Art
14.5.1 Other Artistic Movements Based on Electronics
14.5.2 Alive Art
14.5.3 The Aliver
14.5.4 The "Alive Art Effect"
14.5.5 Example—LEGO Robot Artists
14.6 Conclusions
References
Chapter 15 - Stepping Stones in the Mist
By Paul Brown
15.1 Introduction
15.2 On My Approach as an Artist—A Disclaimer
15.3 Major Influences
15.4 Historical Work—1960s and 1970s
15.5 Early Computer Work
15.6 Recent Work
15.7 Current And Future Directions
15.8 Conclusions
Acknowledgments
References
Chapter 16 - Evolutionary Generation of Faces 409
By Peter J. B. Hancock and Charlie D. Frowd
16.1 Introduction
16.1.1 Eigenfaces
16.1.2 Evolutionary Face Generator System
16.2 Testing
16.2.1 Apparatus
16.2.2 Generation of Face Images
16.2.3 Evolutionary Algorithm
16.2.4 Participants
16.3 Results
16.4 Discussion
16.5 Conclusions
Acknowledgments
References
Chapter 17 - The Escher Evolver: Evolution to the People
By A. E. Eiben, R. Nabuurs, and I. Booij
17.1 Introduction
17.2 The Mathematical System Behind Escher's Tiling
17.3 Evolutionary Algorithm Design
17.3.1 Representation
17.3.2 Ground Shape and Transformation System
17.3.3 Genetic Operators: Mutation and Crossover
17.3.4 Selection Mechanism
17.4 Implementation and The Working of The System
17.4.1 Stand-Alone Version
17.4.2 First Networked Version
17.4.3 Second Networked Version
17.5 Conclusions
Acknowledgments
References
PART V Evolutionary Innovation
Chapter 18 - The Genetic Algorithm as a Discovery Engine: Strange Circuits and New Principles
By Julian F. Miller, Tatiana Kalganova, Natalia Lipnitskaya, and Dominic Job
18.1 Introduction
18.2 The Space of All Representations
18.3 Evolutionary Algorithms That Assemble Electronic Circuits From A Collection of Available Components
18.3.1 Binary Circuit Symbols
18.3.2 Multiple-Valued Circuits
18.4 Results
18.4.1 One-Bit Adder
18.4.2 Two-Bit Adder
18.4.3 Two-Bit Multiplier
18.4.4 Three-Bit Multiplier
18.4.5 Multiple-Valued One-Digit Adder with Carry
18.5 Fingerprinting and Principle Extraction
18.6 Conclusions
References
Chapter 19 - Discovering Novel Fighter Combat Maneuvers: Simulating Test Pilot Creativity
By R. E. Smith, B. A. Dike, B. Ravichandran, A. El-Fallah, and R. K. Mehra
19.1 Introduction
19.2 Fighter Aircraft Maneuvering
19.3 Genetics-Based Machine Learning
19.3.1 Learning Classifier Systems
19.3.2 The LCS Used Here
19.4 "One-Sided Learning" Results
19.5 "Two-Sided Learning" Results
19.6 Differences In Goals And Techniques
19.6.1 Implications of This Goal
19.7 Conclusions
Acknowledgments
References
Chapter 20 - Innovative Antenna Design Using Genetic Algorithms
By Derek S. Linden
20.1 Introduction
20.2 Antenna Basics
20.3 Conventional Designs and Unconventional Applications: The Yagi-Uda Antenna
20.4 Unconventional Designs and Conventional Applications: Crooked-Wire And Treelike Genetic Antennas
20.4.1 The Crooked-Wire Genetic Antenna
20.4.2 Treelike Genetic Antennas
20.5 Conclusions
References
Chapter 21 - Evolutionary Techniques in Physical Robotics
By Jordan B. Pollack, Hod Lipson, Sevan Ficici, Pablo Funes, Greg Hornby, and Richard A. Watson
21.1 Introduction
21.2 Coevolution
21.3 Research Thrusts
21.4 Evolution In Simulation
21.5 Buildable Simulation
21.6 Evolution and Construction of Electromechanical Systems
21.7 Embodied Evolution
21.8 Conclusions
Acknowledgments
References
Chapter 22 - Patenting of Novel Molecules Designed via Evolutionary Search
By Shail Patel, Ian Stott, Manmohan Bhakoo, and Peter Elliott
22.1 Introduction
22.2 Design Cycle
22.3 Hypothesis: Mechanism Of Action
22.4 Experimental Measures And Modeling Techniques
22.4.1 Molecular Modeling
22.4.2 Neural Networks
22.5 Evolution
22.6 Patent Application
22.6.1 Comparing Patent Spaces
22.7 Conclusions
References
Index
Description
The use of evolution for creative problem solving is one of the most exciting and potentially significant areas in computer science today. Evolutionary computation is a way of solving problems, or generating designs, using mechanisms derived from natural evolution. This book concentrates on applying important ideas in evolutionary computation to creative areas, such as art, music, architecture, and design. It shows how human interaction, new representations, and approaches such as open-ended evolution can extend the capabilities of evolutionary computation from optimization of existing solutions to innovation and the generation of entirely new and original solutions.
This book takes a fresh look at creativity, exploring what it is and how the actions of evolution can resemble it. Examples of novel evolved solutions are presented in a variety of creative disciplines. The editors have compiled contributions by leading researchers in each discipline.
If you are a savvy and curious computing professional, a computer-literate artist, musician or designer, or a specialist in evolutionary computation and its applications, you will find this a fascinating survey of the most interesting work being done in the area today.
Key Features
- Explores the use of evolutionary computation to generate novel creations including contemporary melodies, photo-realistic faces, jazz music in collaboration with a human composer, architectural designs, working electronic circuits, novel aircraft maneuvers, two- and three-dimensional art, and original proteins.
- Presents resulting designs in black-and-white and color illustrations.
- Includes a twin-format audio/CD-ROM with evolved music and hands-on activities for the reader, including evolved images, animations, and source-code related to the text.
- Describes in full the methods used so that readers with sufficient skill and interest can replicate the work and extend it.
- Is written for a general computer science audience, providing coherent and unified treatment across multiple disciplines.
Readership
Computer scientists and students, artists, musicians, and specialists in evolutionary computation
Details
- No. of pages:
- 576
- Language:
- English
- Copyright:
- © Morgan Kaufmann 2002
- Published:
- 16th July 2001
- Imprint:
- Morgan Kaufmann
- eBook ISBN:
- 9780080503370
- Hardcover ISBN:
- 9781558606739
Reviews
This volume shows the current state of the art, and the science, of evolutionary creativity. It shows what can--and equally important, what can't--be done at the turn of the new millennium. What will have been achieved by the turn of the next one is anyone's guess. Meanwhile, it's intriguing, it's instructive, it's difficult, and it's fun! --Margaret Boden, author of The Creative Mind, editor of Dimensions of Creativity
Ratings and Reviews
About the Authors
David Corne Author
David W. Corne is a reader in evolutionary computation (EC) at the University of Reading. His early research on evolutionary timetabling (with Peter Ross) resultedin the first freely available and successful EC-based general timetabling programfor educational and other institutions. His later EC work has been in suchareas as DNA pattern mining, promoter modeling, phylogeny, scheduling, layoutdesign, telecommunications, data mining, algorithm comparison issues, and multiobjectiveoptimization. Recent funded work (with Douglas Kell) applies EC directlyto the in vivo optimization of proteins. He is an associate editor of the IEEETransactions on Evolutionary Computation and a founding co-editor of the Journal ofScheduling. Dr. Corne is on the editorial boards of Applied Soft Computing and the InternationalJournal of Systems Science, and he serves on a host of international conferenceprogram committees. Other recent edited books include New Ideas in Optimization(with Marco Dorigo and Fred Glover), Telecommunications Optimization: Heuristic andAdaptive Techniques (with Martin Oates and George Smith), and Creative EvolutionarySystems (with Peter Bentley). He is also a director of Evosolve (United Kingdomregistered charity number 1086384, with Jeanne Lynch-Aird, Paul Marrow, GlenysOates, and Martin Oates), a nonprofit organization that promotes the use of advancedcomputation technologies to enhance the quality of life.
Affiliations and Expertise
University of Reading
Peter Bentley Author
Peter J. Bentley is a Honorary Research Fellow at University College London, known for his research covering all aspects of EC, including multiobjective optimization, constraint handling, artificial immune systems, computational embryology and more, and applied to diverse applications including floor-planning, control, fraud-detection, and music composition. He speaks regularly at international conferences, and is a consultant, convenor, chair and reviewer for workshops, conferences, journals and books on Evolutionary Design and Evolutionary Computation. He has been a guest editor of special issues on Evolutionary Design and Creative Evolutionary Systems in journals, and is the editor of the book Evolutionary Design by Computers (MKP) and is the author of the popular science book, Digital Biology, to publish in May 2001.
Affiliations and Expertise
University College London, U.K.