关键词: 映射, 曲线相似, 离散Fré, chet距离, 多目标优化

Abstract: To resolve the conflicts among the objectives in the multi-objective optimization,the information among the objective values should be objectively mined by using similarity theory.A new multi-objective optimization method based on discrete Fréchet distance was presented.The binary mapping combination of Z-score standardization-Max-value normalization was established,by which Pareto front and corresponding objective function values of ideal solution were mapped into discrete and ordered point chains,then the multiple comparison curves and a reference curve were built.The discrete Fréchet distance was utilized to measure the similar degree between the reference curve and the comparison curves to achieve the multi-objective optimization,and was also deemed as the fitness of GA algorithm to guide the algorithm evolution.Compared with the experiments of the other three algorithms,the optimization solution and various performance indicators obtained by the new algorithm were mostly superior,which proved the feasibility and effectiveness of the new algorithm.

Key words: mapping, similarity between curves, discrete Fréchet distance, multi-objective optimization