Coevolutionary Computation: An Introduction

Agent technology is synonymous with advanced computer software and a growing body of work exists on the use of adaptive techniques within the agent paradigm (e.g. [Kozierok & Maes 1993]). The distributed and/or multi-faceted nature of complex problem domains, such as network routing or process control, has lead to the extension of agent frameworks to multi-agent systems, both within traditional (Distributed) Artificial Intelligence (e.g.[O'Hare & Jennings 1996]) and Machine Learning (e.g. [Weiss 1997]). Here each node/aspect of a problem is under the control of an agent, with many such agents interacting to generate a global solution. The subject of this book is the use of Evolutionary Computation in such systems, termed Coevolutionary Computation.