Evolutionary computing for multidisciplinary optimisation

Multidisciplinary optimisation (MDO) is needed for increasingly complex design problems where system performance characteristics are influenced by more than one discipline, such as the design of an aeroplane. Traditionally, MDO problems were tackled using approximation and decomposition techniques to split a problem into simpler blocks using simple models to give a general picture. These techniques no longer cater for the increasing cost of the design life cycle where a very good and accurate design is preferred at an early stage. Evolutionary computing (EC) techniques have been shown to be a powerful platform for search and optimisation problems involving single and multiple objectives. The MDO community has largely ignored EC so far. The authors introduce the prevailing concepts behind current MDO thinking and then show how EC could be used in MDO. Samples of recent work on MDO using EC are presented. An initial assessment by the authors of the NASA MDO test suite is presented. Examples from the test suite are discussed. Suggestions for future work conclude this paper