Extraction of Emerging Multi-Objective Design Information from COGA Data


Abstract

The paper describes further developments of the interactive evolutionary design concept relating to the emergence of mutually inclusive regions of high performance solutions relating to differing objectives from cluster-oriented genetic algorithm (COGAs) output. These common regions are further defined by the application of clustering algorithms and relevant variable analysis. The multi-objective output of the COGA is then compared to output from a strength Pareto evolutionary algorithm (SPEA-II).