中文摘要：

针对思维进化算法（MEA）群体多样性少,易陷入局部极值的问题,提出了能够学习并进行反思的思维进化算法,同时分析了该算法的机制。设计利用子群体的线性组合实现信息共享,通过适应度值的比较进行反思,运用进化历史最优值控制异化策略,提高了搜索速度,保证了种群的多样性,使得收敛速度和全局收敛性均达到较好平衡。测试函数寻优验证了算法的可行性和高效性。

英文摘要：

In order to add population diversity,mind evolutionary algorithm（MEA） based on consideration was presented.The linear combination of subpopulation was used to realize social information share,and fitness comparing was used consideration.Dissimilation strategy with density control was designed according to history classic individual.The improved algorithm realized optimization process by total optimal characteristics of the whole population,and achieved the best balance between searching velocity and global convergence.Optimization experiments by the testing functions showed the effectiveness and feasibility of the improved MEA.