Lazy CBS: Implicit Conflict-Based Search Using Lazy Clause Generation

Abstract

Conflict-based Search (CBS) is a effective approach to optimal multi-agent path finding. However, performance of CBS approaches degrade rapidly in highly-contended graphs with many agents. One of the reasons this occurs is that CBS does not detect independent subproblems; i.e. it can re-solve the same conflicts between the same pairs of agents up to exponentially many times, each time along a different branch. Constraint programming approaches with nogood learning avoid this kind of duplication of effort by storing nogoods that record the reasons for conflicts. This can exponentially reduce search in constraint programming. In this work, we present Lazy CBS, a new approach to multi-agent pat..