论文部分内容阅读
针对多目标进化算法搜索效率低和收敛性差的问题,提出了基于精英重组的混合多目标进化算法,将多目标优化问题分解为多个单目标优化问题单独求解,并采用基于遗传算法的精英重组策略将多个相异解重组生成唯一的精英解.提出区域化的种群初始化方法,改进局部搜索及群体选择机制,采用以优化子群为核心的分组交叉策略及自适应多位变异算子,并引入基于混沌优化的重启机制,有效克服了精英保存的固有缺陷,以及现有多目标进化算法存在的目标空间解拥挤、收敛慢、易早熟等问题.多目标测试函数的数值仿真和关键步骤的性能分析证明了本文算法的有效性和优越性.
In order to solve the problem of low efficiency and poor convergence of multi-objective evolutionary algorithms, a hybrid multi-objective evolutionary algorithm based on elitist recombination is proposed. The multi-objective optimization problem is decomposed into multiple single-objective optimization problems separately. The elitist recombination based on genetic algorithm In this paper, a unique elitist solution is generated by multiple dissimilar solutions.It is proposed that a regionalized population initialization method be used to improve the local search and group selection mechanism. By using the group crossing strategy with the optimized subgroup as the core and the adaptive multiple mutation operator, And introduces the restart mechanism based on chaos optimization, which effectively overcomes the inherent shortcomings of elitist preservation and the problems of existing target space such as congestion, slow convergence and precocious problems in existing multi-objective evolutionary algorithms.The numerical simulation and key steps of multi-objective test function The performance analysis proves the effectiveness and superiority of the proposed algorithm.