Strategies for the Integration of Evolutionary/Adaptive Search with the Engineering Design Process

Summary

The research concerns the development of evolutionary/adaptive search strategies to enable their successful integration with the conceptual, embodiment and detailed stages of the engineering design process. Global optimisation in relation to engineering design is considered here in its broadest sense, i.e., as a complex, rela tively continuous process that commences during the high risk stages of conceptual design and progresses through the uncertainties of embodiment design to the more deterministic, lower risk stages of detailed design. The objective during the early stages is to identify optimal design direction (i.e., that direction that represents best performance whilst best satisfying many qualitative and quantitative criteria at least risk). During the more deterministic detailed design stages the emphasis is upon minimisation of computational expense whilst identifying optimal design solutions. Appropriate adaptive search integration involves the utilisation of design models of varying detail commensurate with the degree of confidence in available data and project specification. Results from the implementation of co-operative search strategies also involving complementary soft computing techniques are presented and discussed. The development and integration of appropriate strategies is illustrated with examples of real-world application from the mechanical, civil, electronic, aerospace and power system engineering design domains.