Lessons Learned from the Effort to Solve Cooperative Path-Finding Optimally: Reductions to Propositional Satisfiability
Abstract
This paper addresses makespan optimal solving of cooperative path-finding problem (CPF) by translating it to propositional satisfiability (SAT). The task in CPF is to relocate a set of agents to given goal locations so that they do not collide with each other.
Recent findings indicate that a simple direct encoding outperforms the more elaborate encodings based on binary encodings of multi-value state variables. The direct encoding is further improved by a hierarchical build-up that uses auxiliary variables to reduce its size in this work.
The conducted experimental evaluation shown that the simple design of the encoding together with new improvements which reduced its size significantly are key enablers for faster solving of the encoded CPFs than with existing encodings. It has been also shown that the SAT based methods dominates over A* based methods in envi-ronments with high occupancy by agents.