一种改进的粒子群优化算法惯性权值递减策略

An Improved Particle Swarm Optimization Algorithm with the Decreasing Strategy of Inertia Weight

结合粒子群优化(PSO)算法的特点,分析惯性权值的关键性作用。在此基础上提出一种改进的非线性惯性权值递减策略。同时利用两种基准函数对该策略进行测试。实验结果表明,在参数设置均相同的条件下,改进后的权值递减策略在算法迭代初期具有较好的多样性,有利于跳出局部极值,在迭代后期具有更好的全局寻优能力。当维数不变时,随着种群规模以及最大迭代次数的相应增加,改进后的权值递减策略在收敛精度指标上要明显优于对比算法。

Combining with the particle swarm optimization （PSO） algorithm characteristics, it analyzed the key role of inertia weight in this paper. The improved nonlinear inertia weight decreasing strategy was proposed on this basis. Simultaneously the strategy was tested by applying two reference functions. The experimental results showed that the improved strategy is endowed better diversity in the initial stage of the algorithm, it is advantageous to get rid of the affect of the local extremum and enjoys better global optimization ability in the later iterations on the same condition. When the dimension unchanged and with the addition of population size and the maximum number of iterations, the improved nonlinear inertia weight decreasing strategy is superior to the contrast algorithm in the convergence precision.