COMPARISON OF 3 NON-DERIVATIVE OPTIMIZATION METHODS WITH A GENETIC ALGORITHM FOR COMPONENT SELECTION

This paper presents a design method to automate the selection of an optimal set of components for a given layout and set of design specifications. The systems considered in this research are complex, multi-energy domain systems which have transient performance design specifications. Selecting components for such systems requires a wide range of knowledge, including knowledge of system design, modelling, dynamic analysis, computer programming and optimization. The component selection method developed in this work consists of a genetic algorithm and a generalized constraint handling procedure. A genetic algorithm is a flexible, evolutionary, combinatorial optimization technique that lends itself well to problems with discrete solution spaces. This component selection approach is compared with other approaches: a geometric simplex algorithm; an enumerative search; a discrete hill-climbing routine and a random search. The design of a hydraulic lift system was used as the test problem. This method (the genetic algorithm) finds a good design quickly without database ordering or using rules such as an expert system would require.