Multi-agent pathfinding with continuous time

Multi-Agent Pathfinding (MAPF) is the problem of finding paths for multiple agents such that each agent reaches its goal and do not collide. In recent years, variants MAPF have risen in a wide range real-world applications as warehouse management autonomous vehicles. Optimizing common objectives, minimizing sum-of-costs or makespan, computationally intractable, but state-of-the-art algorithms are able to solve optimally problems with dozens agents. However, most assume (1) time discretized into steps (2) duration every action one step. These simplifying assumptions limit applicability raise non-trivial questions how discretize an effective manner. We propose two novel optimal solutions rely on any discretization. particular, our require quantization wait move actions' durations, allowing these durations take value required find solutions. The first algorithm we propose, called Continuous-time Conflict-Based Search (CCBS), draws ideas from Safe Interval Path Planning (SIPP), single-agent pathfinding designed cope dynamic obstacles, (CBS), search-based algorithm. SMT-CCBS builds similar ideas, based different SMT-CBS, which applied SAT Modulo Theory (SMT) problem-solving procedure. CCBS guarantees return minimal sum-of-costs, while makespan. implemented evaluated them grid-based general graphs (roadmaps). results show both can efficiently problems.