粒子群优化算法惯性权重的一种动态调整策略

为了有效地调整粒子群优化算法的全局开拓和局部搜索能力,提出一种基于Logistic模型动态调整惯性权重的粒子群优化算法.该算法在初期保持较大的惯性权重,使其具有较大的全局开拓能力,在进化后期保持较小的惯性权重,有利于局部搜索,加速算法的收敛.通过标准测试函数的仿真实验表明：该调整策略优于线性递减的调整策略,且对于优化多峰值函数具有一定的优越性.

A Dynamic Adjustment Strategy of Inertia Weight in Particle Swarm Optimization

To overcome the disadvantages of particle swarm algorithm, a new particle swarm algorithm was proposed, in which the inertia weight of the particle was adopted dynamically based on the logistic model. It could keep the individual diversity and improved searching ability of global optimum in the pop- ulation at the initial generations. However, the algorithm was gradually stabilized with searching ability of local optimum improving at a later time. Simulation results showed that the proposed algorithm had certain advantages to complex multi-peak optimization problems.