摘要
针对于差分进化算法在高维多峰函数环境下易早熟和迭代收敛速度较慢的问题,通过引入自适应的缩放比例因子的方法,提出了一个基于自适应缩放比例因子的差分进化算法。通过理论推导改进的差分进化算法可以有效提高差分进化算法对于高维多峰函数全局最优值搜索能力和差分进化算法对于高维优化问题的收敛速度,并且通过形式化证明的方法分析了其可以提高着这些性能的具体原因,实验结果表明了理论推导以及对于改进差分进化算法性质分析的正确性。
Aiming to solve the differential evolution problem of prematurity and low iteration speed under high-dimension multimodal function situation, using the adaptive scaling factor, an improved differential evolution algorithm based on adaptive scaling factor is pre sented. Though theoretical derivation, improved differential evolution algorithm is proofed to be better than standard differential evolution algorithm in global optimum searching ability of multi-dimensional and multi-modal function, as well as in iteration speed under high-dimension multi-modal situation, and the reasons of these improvements are also found by the formal proofing. The correct ness of the theoretical derivation and improvement differential evolution algorithm is also verified by experiment.
出处
《计算机工程与设计》
CSCD
北大核心
2014年第1期261-266,共6页
Computer Engineering and Design
基金
国家863高技术研究发展计划基金项目(2013AA01A211)
关键词
差分进化算法
自适应比例因子
高维多峰函数
迭代速度
最优值查找
differential evolution
adaptive scaling factor
high victoria peak function
iteration speed
optimal value searching
