作　　者： ; ; ;

机构地区： 深圳大学数学与计算科学学院

出　　处： 《深圳大学学报（理工版）》 2008年第2期211-215,共5页

摘　　要： 以进化策略算法为框架,提出一种求解连续函数,特别是高维连续函数问题的优化算法——差分进化策略.该算法利用进化策略快速收敛的优点,融入了差分演化算法中具有较强全局搜索能力的变异算子.经数值实验分析表明,差分进化策略在函数优化过程中具有较强稳健性,可提高全局搜索能力,保持快速收敛优势,能用于研究生物进化、机器学习、人工智能、模糊系统及人工神经网络训练等领域. A new algorithm, differential evolutionary strategies （DES）, for the high-dimensional continuous function optimization, was proposed. The proposed algorithm was designed by making use of both the strong global search capability of differential evolutionary strategies and the rapidly converging capability of evolution strategies. Computer simulations were tested on several high-dimensional continuous function optimization problems, and the results indicate that the proposed algorithm improves the efficiency and is much more robust than conventional evolutionary strategies. The proposed algorithm can be used in biological evolution researching, machine learning, artificial intelligence, fuzzy system, artificial neural network training etc. , especially in digital signal processing, data mining and multi-programming.

关 键 词： 差分演化 进化策略 变异算子 全局优化 信号处理 机器学习 人工智能

领　　域： [自动化与计算机技术] [自动化与计算机技术]