一种动态分群的自适应粒子群优化算法

An adaptive particle swarm optimization algorithm with dynamic sub-swarms

粒子群优化算法是模拟鸟类觅食行为思想的随机搜索算法,主要是通过迭代寻找最优解。将粒子随机初始化改进为固定初始化,并将动态分群思想引入粒子群优化算法将整个种群划分为三个子群,根据不同群中粒子的情况自适应地选择惯性权重,以此提高粒子的搜索能力。仿真实验结果表明,该方法大大提高了搜索过程中粒子的多样性,避免粒子陷入局部最优,提高了求解的速度和精度。

Particle swarm optimization algorithm is a random search algorithm simulating birds＇ foraging behavior and thought, which mainly searches the optimal solution by iteration. This paper uses particle fixed initialization instead of particle random initialization and introduces dynamic sub-swarms theory into particle swarm optimization algorithm to divide the whole population into three subgroups in which inertia weight is chosen adaptively according to particle situation of different groups in order to improve the search ability of particle. The simulation experimental results show that the method greatly increases the diversity of particles in the process, keeps particle from trapping in local optimum and improve the solution speed and precision.