摘要: 针对复杂函数的数值积分求解问题,提出一种基于不等距点分割的差分进化算法。该算法在积分区域中选取一些个体,并利用差分进化算法对其进行优化,通过将函数变化较快的区间分得较细,函数变化较慢的区间分得较粗,得到较准确的数值积分。数值计算结果证明,该算法可以快速计算通常意义下任意函数的定积分,并能计算振荡函数的积分。
关键词:
差分进化算法,
数值积分,
不等距点分割,
振荡函数,
适应度
Abstract: An approach based on inequality point segmentation Differential Evolution(DE) algorithm is presented to solve complex function numerical integral. DE is used to optimize the points in the integration interval, and make the rapid-change interval of function smaller or make the slow-change interval of function thicker, so that the more precise result is obtained. Numerical experimental results prove that the algorithm can compute integral for any functions in the usual sense, and compute oscillation function integral.
Key words:
Differential Evolution(DE) algorithm,
numerical integral,
inequality point segmentation,
oscillation function,
fitness
中图分类号:
邓泽喜, 黄飞丹, 刘晓冀. 一种求解数值积分问题的差分进化算法[J]. 计算机工程, 2011, 37(20): 206-207.
DENG Ze-Chi, HUANG Fei-Dan, LIU Xiao-Ji. Differential Evolution Algorithm for Numerical Integral Problem[J]. Computer Engineering, 2011, 37(20): 206-207.