Abstract: Indicator Based Evolutionary Algorithm(IBEA) is an excellent multi-objective optimism algorithm. IBEA has outstanding performance for convergence, but it has poor performance for maintaining diversity on some test problems. This paper studies the fitness assignment strategy of IBEA and improves it for its inferiors. Test and comparison indicate that the improved IBEA maintains the advantage of the original IBEA and shows better performance for maintaining diversity than original IBEA.

Key words: multi-objective optimism, evolutionary algorithm, quality indicator

摘要： 基于指标的进化算法(IBEA)是一个出色的多目标优化算法。IBEA具有良好的收敛性，但在保持解的多样性方面对于某些问题却表现较差。对IBEA进行研究，分析其适应度分配原理，针对其缺点进行改进，并将IBEA与其他2个算法进行了测试比较。测试结果表明改进后的IBEA在保持了原算法优点的情况下使其在解的多样性方面有了较大改观。

关键词: 多目标优化, 进化算法, 性能指标

CLC Number: 

• TP301.6