摘要
针对动态环境中的种群多样性问题,提出一种保持种群多样性的双子群粒子群优化算法。将群搜索算法中的游走者思想引入到粒子群优化算法中,基于群体多样性,子种群B采用不同的方法更新速度和位置,子种群A和子种群B交换最优信息,扩展种群的搜索范围,增强整个群体的多样性水平。将改进的算法应用于复杂变化的抛物线函数和群体动画的跟随效果中,结果表明该算法在动态环境中的有效性,并能够真实模拟群体跟随行为。
A double population particle swarm optimization with adaptive diversity preservation is proposed considering the population diversity in dynamic environment. The ranger idea of group search is introduced to particle swarm optimization, where subswarm B updates its speeds and positions with different methods according to the diversity of particle swarm and subswarm A and B exchange their optima. These mechanisms extend the search range and improve the swarm diversity. The scheme is tested on benchmark functions with dynamic complex changes and the simulation results show the proposed algorithm is effective in dynamic environments. It is also used to simulate group following behavior.
出处
《计算机工程》
CAS
CSCD
2012年第16期167-169,173,共4页
Computer Engineering
基金
山东省高等学校科技计划基金资助项目(J10LG08)
山东省优秀中青年科学家科研奖励基金资助项目(BS2010DX033)
关键词
动态粒子群优化
多样性
双种群
群搜索
群体动画
dynamic particle swarm optimization
diversity
double population
group search
group animation
