The article discusses the targeted parallelization of conflict-based search for Multi-Robot Path Planning (MRPP). The authors have introduced a method that utilizes a decentralized parallel algorithm to explore multiple branches in instances where robots are densely packed. On the other hand, for large instances with sparse robot interactions, the method prioritizes node expansion and conflict resolution. This multi-threaded approach significantly improves the success rate or runtime over baseline serial methods. The study provides further insights into MRPP and presents a promising path for improving solution quality and computational efficiency through parallel algorithmic strategies.
Publication date: 20 Feb 2024
Project Page: Not provided
Paper: https://arxiv.org/pdf/2402.11768
