A Mobile Robot Path Planning Method Based on Dynamic Multipopulation Particle Swarm Optimization

Abstract

To overcome the limitations of particle swarm optimization (PSO) in mobile robot path planning, including issues such as premature convergence and sensitivity to local optima, this study proposes a novel approach, dynamic multipopulation particle swarm optimization (DMPSO). First, the multipopulation particle swarm optimization (MPSO) framework is extended by introducing a dynamic multipopulation strategy that adjusts the number of subpopulations in real-time. This strategy is designed to enhance the algorithm’s local search capabilities and accelerate its convergence. Second, the inertia weights and learning factors within the algorithm are refined to achieve a balance between global exploration and local exploitation. Furthermore, an initialization strategy based on fitness variance is developed to improve population diversity, mitigate premature convergence, and enhance the algorithm’s ability to locate global optima. Lastly, a positive feedback acceleration factor is introduced to optimize particle positions, thereby improving local search capabilities and accelerating convergence. Simulation experiments have validated that DMPSO offers improved exploration capabilities, enhanced search precision, and a more rapid convergence rate. In comparison to PSO, DMPSO reduces the path length by 3% and decreases the number of convergence iterations by 17%.