高维多峰函数的量子行为粒子群优化算法改进研究  

Improvement of quantum-behaved particle swarm optimization algorithm for high-dimensional and multi-modal functions

作　　者：田瑾[1] 

机构地区：[1]首都经济贸易大学信息学院,北京100070

出　　处：《控制与决策》2016年第11期1967-1972,共6页Control and Decision

基　　金：北京自然科学基金项目(9142003)

摘　　要：针对群智能优化算法求解高维多峰函数难以优化粒子每一维和易陷入局部极值点问题,在分析量子行为粒子群优化(QPSO)算法机理的基础上,对QPSO算法进行改进,采取前后代粒子逐维对比优化,并构造一种新的调控收缩扩张系数函数.实验结果表明,改进算法在收敛精度和收敛速度上明显优于QPSO算法,具有很强的避免陷入局部最优的能力,非常适合求解高维、多峰优化问题.For the swarm intelligence optimization algorithm of solving high-dimensional and multi-modal functions, it is difficult to optimize the particles for each dimension, and it is easy to fall into the local extreme point. On the basis of analyzing the mechanism of quantum-behaved particle swarm optimization（QPSO） algorithm, the QPSO algorithm is improved. Each dimension of the previous generation particle is compared with the later generation to optimize, and a new control function of the contraction-expansion coefficient is constructed. The experimental results show that the improved algorithm significantly outperforms the QPSO algorithm in the convergence accuracy and convergence rate. Specifically, it is of strong ability to avoid falling into the local optimum, and is very suitable for solving high-dimensional and multi-modal optimization problems.

关 键 词：量子行为粒子群优化 高维多峰函数 收缩扩张系数 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]