摘要
为了提高入侵性杂草优化算法(IWO)在搜索深度上的不足,使算法在处理连续性问题时具有更好的全局收敛性,根据杂草算法在搜索上的广度和粒子群算法(PSO)在搜索上的深度,提出了一种改进的IWOPSO混合算法。该算法在子代扩散中以PSO算法中的位置、速度公式代替了杂草算法中的正态分布方式,引入一个随机数对新的子代个体进一步正态分布,提高了算法后期的局部搜索能力,使算法收敛到更好的全局最优解。利用5个benchmark函数测试算法的寻优能力,仿真结果表明,无论对于多峰还是单峰函数,低维还是高维函数,IWOPSO算法的收敛速度和最优解都要优于标准IWO和PSO算法。
In order to improve the invasive weed optimization algorithm （IWO） deficiency in the search depth, and make the global convergence of the algorithm better in processing continuous problem, according to weed algorithm in search of the breadth and particle swam] optimization （PSO） in the search depth, this paper proposes an improved IWOPSO hybrid algorithm. In the progeny of diffusion, this algorithm insteads the normal distribution of weed with position and velocity formula of PSO algorithm, and a random number of offspring to normal distribution is introduced to enhance the local s.earehing ability of the later algorithm, and converge to the global optimal better solutions. This paper uses 5 benchmark function to test the searching capability of the algorithm. The simulation results show that, whether for muhimodal and unimodal function, low dimensional and high-dimensional function, the convergence speed of IWOPSO algorithm and the optimal solution is better than the standard IWO algorithm and PSO algorithm.
出处
《微型机与应用》
2014年第24期66-68,共3页
Microcomputer & Its Applications
基金
芬兰科学院基金(135225)
关键词
入侵性杂草优化
混合
正态分布
全局优化
invasive weed optimization
hybrid
normal distribution
global optimization
