一种快速自适应粒子群算法

A Quickly Adaptive Particle Swarm Optimization Algorithm

为了解决标准粒子群优化算法(SPSO)不能适应复杂非线性优化过程的问题,提出了一种动态改变惯性权重的快速自适应粒子群优化算法(QAPSO),直接利用群粒子的位置分布情况控制粒子飞行的惯性权重,借助于个体最优位置和全局最优位置的平均作用避免粒子陷入局部最优。通过多个基准函数仿真结果表明,在不引入额外设计及增加实现复杂度的前提下,相对于SPOS等经典算法,QAPSO在收敛速度、最优解精度等方面获得了大幅提升,尤其对于多峰函数效果更明显。

A quickly adaptive particle swarm optimization algorithm （QAPSO） with dynamically changing inertia weight was presented to solve the problem that the Standard Particle Swarm Optimization algorithm （SPSO） cannot adapt to the complex and nonlinear optimization process. The QAPSO evaluates the population distribution, and automatic control inertia weight. By using the mean of the optimal individual position and the global optimal position, the best global particle can jump out of the lo cal optima. The QAPSO has comprehensively been evaluated based on benchmark functions. Results show that QAPSO sub stantially enhances the performance of paradigm such as the SPSO in terms of convergence speed, solution accuracy without in- troducing an additional design or implementation complexity, especially for multimodal functions.