Multi-agent path planning based on meta-agent conflict merge search algorithm

Multi-agent path planning is one of the research difficulties in the field of agents. Conflict-based search (CBS) algorithm will branch Con-straintTree (CT) for all detected conflicts, resulting in poor agent search path efficiency. Therefore, a meta-agent conflict merge search algorithm is proposed. Firstly, a new conflict detection strategy is proposed, which stores all agent paths in the form of multi-valued decision graph. Then, hierarchical traversal is used to detect conflicts between agents, and non-major conflicts are merged to reduce CT branch nodes. For major conflicts, CT branch constraints are carried out according to the order of traversal, so that conflicts that are detected first are resolved first. The experimental results show that the proposed algorithm can effectively improve the solving efficiency and the success rate of obstacle avoidance.