关键词: 多目标优化进化算法, 目标侧重度, 目标间距, 多样性

Abstract:

A new method of multi-objective optimization based on the customer requiremnents was designed.In order to improve the efficiency of the traditional algorithm, the concept of object focus degree was proposed,so that the evolutionary algorithm convergence was of selective.There are several limitations of the existing evolutionary algorithms, such as easily falling into local optimal solution and uneven solution distribution. According to these weaknesses,a concept of object spacing matrix was proposed.The
new method can avoid the earlier algorithm convergence and make sure
the diversity of the optimal solution.The proposed method was simulated based on
MATLAB7.0.The simulation results indicate that the proposed algorithm is feasible,
and has better reliability and superiority than other two popular evolutionary algorithms(NSGA-2 and SPEA-2). 

Key words: multi-objective optimization algorithm, object focus degree, object space, diversity