An empirical comparison of the hardness of multi-agent path finding under the makespan and the sum of costs objectives

Abstract

In the multi-agent path finding (MAPF) the task is to find non-conflicting paths for multiple agents. Recently, existing makespan optimal SAT-based solvers for MAPF have been modified for the sum-of-costs objective. In this paper, we empirically compare the hardness of solving MAPF with SAT-based and search-based solvers under the makespan and the sum-of-costs objectives in a number of domains. The experimental evaluation shows that MAPF under the makespan objective is easier across all the tested solvers and domains.