Multi-Agent Path Finding for Large Agents

Abstract

Multi-Agent Path Finding (MAPF) has been widely studied in the AI community. For example, Conflict-Based Search (CBS) is a state-of-the-art MAPF algorithm based on a twolevel tree-search. However, previous MAPF algorithms assume that an agent occupies only a single location at any given time, e.g., a single cell in a grid. This limits their applicability in many real-world domains that have geometric agents in lieu of point agents. Geometric agents are referred to as “large” agents because they can occupy multiple points at the same time. In this paper, we formalize and study LAMAPF, i.e., MAPF for large agents. We first show how CBS can be adapted to solve LA-MAPF. We then present a generalized version of CBS, called Multi-Constraint CBS (MCCBS), that adds multiple constraints (instead of one constraint) for an agent when it generates a high-level search node. We introduce three different approaches to choose such constraints as well as an approach to compute admissible heuristics for the high-level search. Experimental results show that all MC-CBS variants outperform CBS by up to three orders of magnitude in terms of runtime. The best variant also outperforms EPEA* (a state-of-the-art A*-based MAPF solver) in all cases and MDD-SAT (a state-of-the-art reduction-based MAPF solver) in some cases.

Generalized and Sub-Optimal Bipartite Constraints for Conflict-Based Search

“…Agents have a physical shape such as spheres, polygons or polygonal meshes which are situated relative to their reference point (Li et al 2019b). Agents move along edges (v(s n ), v(s n+1 )) ∈ E -an agent begins with its reference point at v(s n ) at time t(s n ) and its reference point follows a straight, constant velocity motion vector in metric space, ending at v(s n+1 ) at time t(s n+1 ).…”

“…Several optimal enhancements for CBS in the classic problem have been published, (Boyarski et al 2015b;Felner et al 2018;Li et al 2019a;Gange, Harabor, and Stuckey 2019). CBS has also been extended for several new problem variants: large agents (MC-CBS) (Li et al 2019b), robustness (Atzmon et al 2018), road maps (Hönig et al 2018), convoys (Thomas, Deodhare, and Murty 2015), trains (Atzmon, Diei, and Rave 2019) and deadlines (Ma et al 2018). Sub-optimal algorithms have also been published (Barer et al 2014;Cohen and Koenig 2016;Walker, Chan, and Sturtevant 2017;Cohen et al 2018).…”

“…CBS with time-annotated bicliques (CBS+TAB) utilizes bipartite reduction (BR) and is a generalization of multiconstraint CBS (MC-CBS) (Li et al 2019b) which allows constraint sets C, to be used with CT nodes. MC-CBS was applied to unit-cost domains with large agents which was shown to reduce the number of CT node expansions because more conflicts are resolved per CT node.…”

“…The goal of multi-agent pathfinding (MAPF) is to move multiple agents to their respective goal states while avoiding conflicts between agents. Conflict-based search (CBS) (Sharon et al 2015), a state-of-the-art algorithm for MAPF, is designed around detecting and resolving conflicts between agents by systematically adding constraints to agents and has been used for many variants of the MAPF problem (Atzmon et al 2018;Li et al 2019b;Ma et al 2018;Li et al 2019a;Thomas, Deodhare, and Murty 2015;Hönig et al 2018). Most notably, CBS has been adapted for nonunit cost and continuous time domains (Andreychuk et al 2019;Cohen et al 2019).…”

Multi-agent Path Finding with Capacity Constraints

“…In order to reflect various aspects of real-life applications variants of MAPF have been introduced such as those considering kinematic constraints [8], large agents [11], or deadlines [14] -see [13] for more variants.…”