Implementation of Dynamic Path Planning Method Driven by Particle Swarm Optimization Algorithm in Intelligent Decision-Making System

Facing the limitations of traditional static path planning in dynamic environments regarding real-time obstacle avoidance and multi-task coordination, this paper proposes a PSO-driven dynamic path planning method. Firstly, a dynamic environment modeling module fuses multi-sensor data for real-time perception using probabilistic occupancy grids. Secondly, an enhanced PSO path generator employs adaptive inertia weights (adjusted by environmental complexity) and dynamic neighborhood topology to optimize search efficiency and responsiveness. Finally, a rolling horizon optimization strategy and priority-based task scheduler enable real-time path updates and seamless task switching. Experimental results demonstrate significant advantages: in high obstacle density (0.8/m²), the method achieved a 95% replanning success rate (vs. A*'s 60%, Dijkstra's 66%) with a stable response time of 0.52±0.05s; under multi-task switching, it maintained a 96.8% path adjustment success rate with an average response delay of only 452ms, significantly outperforming baseline algorithms.