A comparison of evolutionary-based strategies for mixed discrete multilevel design problems

摘要

This paper presents a comparison of evolutionary strategies which employ individual mutation schemes for different types of decision variables in optimal design and control.. problems. The work utilises the GAANT[9] algorithm as an improvement onprevious work involving a dual-agent GA integrated with a nuclear power station whole plant design problem. The objective of the algorithm is to maintain diversity across both discrete and continuous variables. The algorithm is important during thepreliminary design stages of industrial design problems with a limited number of discrete paths and heavy constraint. Particularly a nuclear power station re-design problem has been studied in depth.