The focus of this workshop is to discuss the integration of evolutionary computation (genetic algorithms, genetic programming, evolution strategies, etc.) in design. Particular attention will be paid to design representations suitable for evolution (grammars, L-systems, spatial partitioning) and how factors such as epistasis, or premature convergence of structure before detail, influence the evolution of good designs using such representations. In addition, attention will be paid to the broadening of the definition of evolutionary design, to encompass the ideas and techniques used in Evolutionary Art and Artificial Life.

Questions to be answered will include: should we continue to use evolutionary computation more as generative tools, instead of simply optimizers? How can we convince designers of the fact that an unpredictable, unexplainable, stochastic method is of use to them? Can we use ideas from other fields, including biology, to increase the capabilities of our computational models? Is there a future in using Evolutionary Computation in design, or will its limitations (e.g. being unable to backtrack or 'undo' stages of evolution) ultimately prevent us from tackling unsimplified real-world problems?

Convenor: Peter Bentley, University College London

Contact person: Peter Bentley, p.bentley@cs.ucl.ac.uk

Advisory Committee: Paul Coates, University of East London, UK; Phil Husbands, Sussex University, UK; Ian Parmee, Plymouth University, UK; Gordon Robinson, Southampton University, UK; Mike Rosenman, University of Sydney, Australia

Conference Details

AID'98 Workshops