无惯性自适应精英变异反向粒子群优化算法

发布时间：2018-04-11 15:36

本文选题：无惯性速度更新式 + 一般性反向学习　；参考：《通信学报》2017年08期

【摘要】：为解决反向粒子群优化算法计算开销大、易陷入局部最优的不足,提出一种无惯性的自适应精英变异反向粒子群优化算法(NOPSO)。NOPSO算法在反向学习方法的基础上,广泛获取环境信息,提出一种无惯性的速度(NIV)更新式来引导粒子飞行轨迹,从而有效加快算法的收敛过程。同时,为避免早熟现象的发生,引入了自适应精英变异策略(AEM),该策略在扩大种群搜索范围的同时,帮助粒子跳出局部最优。NIV与AEM这2种机制的结合,有效增加了种群多样性,平衡了反向粒子群算法中探索与开发的矛盾。实验结果表明,与主流反向粒子群优化算法相比,NOPSO算法无论是在计算精度还是计算开销上均具有较强的竞争能力。
[Abstract]:In order to solve the problem that the inverse particle swarm optimization (RPSO) algorithm has a large computational overhead and is easy to fall into the local optimum, an inertial adaptive elite mutation inverse PSO algorithm (NOPSO.NOPSO) is proposed. Based on the reverse learning method, the reverse PSO algorithm is widely used to obtain environmental information.A new type of inertial velocity nv (NIV) is proposed to guide the particle trajectory, thus speeding up the convergence of the algorithm.At the same time, in order to avoid precocious phenomenon, an adaptive elite mutation strategy is introduced. This strategy not only expands the search range of population, but also helps particles jump out of the combination of local optimum. NIV and AEM, and effectively increases population diversity.The contradiction between exploration and development in reverse particle swarm optimization is balanced.The experimental results show that the NoPSO algorithm is more competitive than the mainstream inverse particle swarm optimization algorithm in terms of computational accuracy and computational overhead.
【作者单位】：武汉大学计算机学院;江西理工大学应用科学学院;华南农业大学信息学院;
【基金】：国家自然科学基金资助项目(No.61170305,No.60873114) 江西省教育厅科学技术研究基金资助项目(No.GJJ161568,No.GJJ151521)~~
【分类号】：TP18

【相似文献】