基于控制理论的微粒群算法分析与改进

Analysis and Improvement About Particle Swarm Optimization Based on Linear Control Theory

利用控制理论对基本微粒群算法、标准微粒群算法及带有收缩因子的微粒群算法进行了详细的分析,结果表明它们或为一个积分环节与两个惯性环节组成的系统;或为三个惯性环节组成的系统.为了提高算法效率,建立了由积分环节与震荡环节组成的系统所对应的改进微粒群算法,并讨论了参数的选择方法.仿真实例证明了该算法的正确性与有效性.

Using control theory, three different particle swarm optimization （PSO）, such as basic PSO, standard PSO and PSO with constriction factor were analyzed in detail, the results showed that the evolution equations were composed of both integral elements and inertia elements or inertia elements only. To improve the calculation efficiency, the inertia elements in evolution equations were modified to the oscillation elements and the corresponding algorithm and the coefficients selection principles were discussed. The optimization computing of some examples was made to show the new algorithm has better global search capacity and rapid convergence rate.