The assumption of synchronized actions in Multi-Agent Path Finding (MAPF) has long constrained its real-world applicability. While Continuous-time Conflict-Based Search (CCBS) emerged to handle asynchronous actions (MAPF-AA), it faltered due to an uncountably infinite state space. This paper introduces Conflict-Based Search with Asynchronous Actions (CBS-AA), a novel approach that tackles this fundamental challenge head-on.
Bypassing Infinite State Spaces for MAPF-AA Completeness
The core innovation lies in CBS-AA’s ability to circumvent the theoretical incompleteness plaguing CCBS. By reformulating the problem, CBS-AA achieves guaranteed completeness and optimality for MAPF-AA. This breakthrough is critical for deploying MAPF in dynamic environments where precise timing is not guaranteed, directly addressing a key limitation identified in prior work. The researchers' method provides a robust foundation for reliable asynchronous multi-agent coordination.
