DEVELOPMENT OF A GENETIC ALGORITHM BASED SEARCH STRATEGY SUITED FOR DESIGN OPTIMISATION OF INTERNAL COMBUSTION ENGINES

Abstract

Engine design optimisation is a multi-objective, multi-domain problem in a discontinuous design space. The state of the
art of optimisation techniques shows that only methods of direct and adaptive search are appropriate for this type of
problem. These include, adaptive random search, simulated annealing, evolution strategies and genetic algorithms. Of
these methods, the genetic algorithms have been shown to be the most suited for the optimisation of multi-modal response
functions in a discontinuous design space. This paper considers the important characteristics of genetic algorithms and
their adaptation for use in parametric design optimisation of internal combustion engines. In order to verify the basic
functionality of the proposed optimisation strategy, a genetic algorithm based, optimisation software was developed and
tested on a number of analytical functions, selected from optimisation literature, with satisfactory results.