Evolutionary Computing Strategies for Preliminary Design Search and Exploration

Abstract

This overview paper illustrates the manner in which appropriate strategies utilizing evolutionary computing and other computational intelligence technologies can result in their successful integration with preliminary design search and exploration processes. Although the various algorithms are still largely perceived as optimisers with specific areas of application, a major generic potential is apparent when the technology is appropriately utilised within a design search and exploration environment. This is optimisation in the broadest sense of the term where the techniques address problems encountered during the early stages of design and the primary task is to identify best direction.

Various developed evolutionary and adaptive search strategies taht address generic problems across the design process are introduced. For instance, the early identification og high-performance regions of a complex preliminary design space, whole system design / mixed integer optimisation, constraint satisfaction and optimisation, the handling of multiple quantitative and qualitative criteria, computational expense and response curve generation. The final area discussed relates to human interaction aspects and the utilisation of evoluationary systems to provide optimal design information to support early decision-making processes.