Computer Engineering and Applications ›› 2012, Vol. 48 ›› Issue (8): 21-23.

• 博士论坛 • Previous Articles     Next Articles

Note on drift analysis

ZHOU Yuren, LAI Xinsheng   

  1. School of Computer Science and Engineering, South China University of Technology, Guangzhou 510640, China
  • Received:1900-01-01 Revised:1900-01-01 Online:2012-03-11 Published:2012-03-11

关于漂移分析的注记

周育人,赖鑫生   

  1. 华南理工大学 计算机科学与工程学院,广州 510640

Abstract: In spite of many successful applications of evolutionary algorithms in various complicated optimization problems, the rigorous theoretical analyses of EAs are still in their infant period. The running time of an evolutionary algorithm for combination optimization problem is an important topic in theoretical study of EAs. This paper discusses the relation between drift analysis and evolutionary algorithms’ time complexity. It gives a new proof of the drift theorem. It also uses drift theorem to estimate the expected runtime of global (1+1) for linear function. The result is a better understanding of the performance of evolutionary algorithms.

Key words: evolutionary algorithms, time complexity, drift analysis

摘要: 进化算法成功应用于求解各种复杂优化问题,其理论研究尚处于初级阶段。时间复杂性分析可以估计算法的平均运行时间,是进化算法理论研究中的重要方向和有力工具。讨论了漂移分析和进化算法时间复杂性的关系,利用吸收马尔科夫链给出漂移定理的一个新的证明;用一步平均漂移估计算法计算时间,得到了线性函数进化算法时间复杂度的一个一般性的结果。这些结果有助于更好地理解进化算法的工作原理和性能。

关键词: 进化算法, 时间复杂性, 漂移分析