Design and Experimental Study of a Parallel RHCR Algorithm for Real-Time Multi-Agent Scheduling

With the continuous expansion of multi-agent systems, path planning is increasingly challenged by high agent density, dynamic task arrivals, and real-time responsiveness. The Rolling Horizon Collision Resolution (RHCR) algorithm demonstrates strong adaptability in dynamic scheduling scenarios due to its local replanning strategy and path stability. However, its current serial architecture limits scalability and computational efficiency in large-scale environments. To address this bottleneck, this paper proposes a parallel optimization framework for RHCR, incorporating multi-threaded path replanning and parallel conflict detection to enhance overall planning performance and system scalability. A modular parallel architecture is designed and implemented, and extensive simulation experiments are conducted. The results show that the proposed algorithm significantly outperforms traditional approaches in terms of planning time, path quality, and system throughput, demonstrating strong practical value and promising potential for future research.