基于双变异策略的自适应骨架差分进化算法

发布时间：2018-04-21 02:15

本文选题：差分进化 + 骨架算法　；参考：《通信学报》2017年08期

【摘要】：骨架差分进化算法能够较好规避差分进化算法控制参数和变异策略选择问题。针对基于双变异策略的经典骨架差分算法(MGBDE)没有根据个体进化差异选择适合的变异策略和考虑早熟收敛的问题,提出一种改进算法。该算法引入变异策略选择因子,并借鉴自适应差分进化算法的设计思想,将选择因子随个体共同参与进化,使个体执行当前最为适合的变异策略,克服原始算法进化过程的盲目性,同时选择因子的动态自适应特性保持了骨架算法近似无参数的优点;该算法加入停滞扰动策略,降低陷入局部最优的风险。采用18个标准测试函数进行实验,结果表明,新算法在收敛精度、收敛速度和顽健性上整体优于多种同类骨架算法以及知名的差分进化算法。
[Abstract]:Skeleton differential evolution algorithm can avoid the problem of parameter selection and mutation strategy selection in differential evolution algorithm. To solve the problem that the classical skeleton difference algorithm based on double mutation strategy (MGBDE) does not select suitable mutation strategy according to individual evolution difference and consider the problem of premature convergence, an improved algorithm is proposed. In this algorithm, the selection factor of mutation strategy is introduced, and the design idea of adaptive differential evolution algorithm is used to make the selection factor take part in the evolution with the individual, so that the individual can carry out the most suitable mutation strategy. It overcomes the blindness of the evolution process of the original algorithm and maintains the advantage of the skeleton algorithm which is approximately parameterless because of the dynamic adaptive feature of the selection factor. The algorithm adds the stagnation perturbation strategy to reduce the risk of falling into the local optimum. By using 18 standard test functions, the experimental results show that the new algorithm is superior to many similar skeleton algorithms and known differential evolution algorithms in convergence accuracy, convergence speed and robustness.
【作者单位】：解放军信息工程大学;中国人民解放军75741部队;中国人民解放军国防信息学院;
【基金】：国家自然科学基金资助项目(No.61309018)~~
【分类号】：TP18

【相似文献】