论文部分内容阅读
收缩—扩张(CE)系数是量子粒子群优化算法(QPSO)需要人工设定的最核心参数,如何选择该参数成为一个重要的问题。为寻找更为有效的CE系数控制方法,根据CE系数递减思想,提出了一种凸凹性可变的指数型非线性下降CE系数控制策略。采用Sphere、Rastrigrin、Griewank和Ackley四种典型的单峰与多峰标准测试函数研究了CE系数的不同控制策略以及不同初始值对量子粒子群优化算法收敛精度与收敛速度的影响,并与线性下降CE系数及固定CE系数两种控制策略进行了对比分析,得出了CE系数控制策略选择的一般性指导准则,为量子粒子群优化算法的应用提供依据。
The constriction-expansion (CE) coefficient is the core parameter of Quantum-Particle Swarm Optimization (QPSO) that need to be manually set. How to choose this parameter becomes an important issue. In order to find a more effective control method of CE coefficient, a control strategy of exponential nonlinear decline CE coefficient with convex-concave variable is proposed according to the idea of decreasing CE coefficient. The effects of different control strategies and different initial values on the convergence accuracy and convergence speed of the quantum particle swarm optimization algorithm were studied by using four typical single-peak and multi-peak standard test functions of Sphere, Rastrigrin, Griewank and Ackley. CE coefficient and fixed CE coefficient were compared and analyzed. The general guideline for CE coefficient control strategy selection was obtained, which provided the basis for the application of quantum particle swarm optimization algorithm.