基于动态惯性权重与莱维飞行的改进粒子群算法

Improved Particle Swarm Algorithm Based on Dynamic Inertia Weight and Levi’s Flight

摘要: 针对粒子群算法在解决高维优化问题时存在的易陷入局部最优的问题，文中引入了具有下降趋势的动态变化因子，提出了一种融合莱维飞行机制的带有动态惯性权重的改进粒子群优化算法(WLPSO)。该算法在位置更新的惯性权重部分引入了一个自适应值，来平衡局部搜索与全局探索能力，同时引入莱维飞行策略增强算法跳出局部最优的能力。在搜索过程中，莱维飞行的长步长特性有效引导粒子群在最优解空间进行高效探索，显著提升了收敛速度。改进后的粒子群算法与其他3种优化算法在9个经典测试函数上进行仿真实验，结果表明改进的粒子群算法在收敛速度和收敛精度方面的综合表现都优于其它算法。

Abstract: Aiming at the problem that the particle swarm optimization algorithm is prone to fall into local optimum when solving high-dimensional optimization problems, a dynamic change factor with a downward trend is introduced in this paper, and an improved particle swarm optimization algorithm with dynamic inertia weight (WLPSO) integrating the Levi flight mechanism is proposed. This algorithm introduces an adaptive perturbation value in the inertia weight part of position update to balance the local search and global exploration capabilities. Meanwhile, it introduces the Levi flight strategy to enhance the algorithm’s ability to escape from local optima. During the search process, the long step size characteristic of Levi’s flight effectively guides the particle swarm to conduct efficient exploration in the optimal solution space, significantly improving the convergence speed. The improved particle swarm optimization algorithm was subjected to simulation experiments with three other optimization algorithms on nine classical test functions. The results show that the comprehensive performance of the improved algorithm in terms of convergence speed and convergence accuracy is superior to that of the other algorithms.