Ron Zvi Stern

Several variants of A* have been recently proposed for finding optimal solutions for the multi-agent pathfinding (MAPF) problem. However, these variants have not been deeply compared either quantitatively or qualitatively. In this paper we aim to fill this gap. In addition to obtaining a deeper understanding of the existing algorithms, we describe in detail the application of the new enhanced partial-expansion technique to MAPF and show how pattern databases can be applied on top of this technique.