计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (33): 43-45.DOI: 10.3778/j.issn.1002-8331.2008.33.013

• 理论研究 • 上一篇    下一篇

多峰值全局优化的混合算法

刘文强1,张晓伟2   

  1. 1.西京学院 数学教研室,西安 710123
    2.西安电子科技大学 应用数学系,西安 710071
  • 收稿日期:2008-06-16 修回日期:2008-09-09 出版日期:2008-11-21 发布日期:2008-11-21
  • 通讯作者: 刘文强

Hybrid algorithm for multi-peak global optimization

LIU Wen-qiang1,ZHANG Xiao-wei2   

  1. 1.Teaching and Research Section of Mathematics,Xijing University,Xi’an 710123,China
    2.Department of Applied Mathematics,Xidian University,Xi’an 710071,China
  • Received:2008-06-16 Revised:2008-09-09 Online:2008-11-21 Published:2008-11-21
  • Contact: LIU Wen-qiang

摘要: 根据遗传算法在较小的可行区域内一般有较好的求解结果这一事实,提出了一种混合算法。该算法先利用区间算法求解全局优化问题来得到包含所有最优解的小区间,随后运用遗传算法进行后续过程。算法能够有效缩小一个较大的可行区域空间,提供高适应值的初始种群,求出多峰值问题的全部最优解,提高算法的求解精度同时避免陷入局部最优。最后数值实验说明了算法的有效性。

关键词: 遗传算法, 区间算法, 全局优化

Abstract: According to the fact that Genetic Algorithm can give generally the better solution in the smaller feasible domain,a hybrid algorithm is presented.The proposed algorithm firstly uses the interval method to obtain all small intervals for the global optimization,then employs the Genetic Algorithm to execute the later process.The hybrid algorithm can reduce efficiently the larger feasible domain,provides the initial population with high fitness,gives all optima of multi-peak optimization,improves the accuracy and avoid falling into local optimum.Finally,numerical experiments show that the algorithm works efficiently.

Key words: Genetic Algorithm, interval method, global optimization