基于高斯函数递减惯性权重的粒子群优化算法

为了有效地平衡粒子群优化算法的全局搜索和局部搜索能力,提出了一种基于高斯函数递减惯性权重的粒子群优化(GDIWPSO)算法。此算法利用高斯函数的分布性、局部性等特点,实现了对惯性权重的非线性调整。仿真过程中,首先对测试函数优化以确定惯性权重的递减方式;然后比较了该算法与权重线性递减、凸函数递减、凹函数递减的粒子群算法优化不同测试函数的性能;最后结果表明,提出的算法在搜索能力、收敛速度及执行效率等方面均有很大提高。

Particle swarm optimization algorithms with decreasinginertia weight based on Gaussian function

To efficiently balance the global search and local search ability, this paper presented a particle swarm optimizationPSO algorithm with decreasing inertia weight based on Gaussian funtionGDIWPSO, this algorithm took advantage of the distribution and locality property of Gaussian function to implement nonlinear inertia weight adjustment. In simulation experiment, optimizing the benchmark function to determine the strategy of decreasing inertia weight and comparing the performance with weight of linear decreasing, convex function decreasing and concave function decreasing. The stimulation results show that the proposed PSO algorithm has better improvement in search ability, convergence rate and computation efficiency.