AAAI

Association for the Advancement of Artificial Intelligence

Home > Proceedings / SOCS-22 Volume 15

Multi-agent Pathfinding on Large Maps Using Graph Pruning: This Way or That Way? (Extended Abstract)

Authors

Jiří Švancara,Philipp Obermeier,Matej Husár,Roman Barták,Torsten Schaub

Charles University, Prague, Czech Republic,University of Potsdam, Potsdam, Germany Potassco Solutions, Potsdam, Germany,Charles University, Prague, Czech Republic,Charles University, Prague, Czech Republic,University of Potsdam, Potsdam, Germany Potassco Solutions, Potsdam, Germany


Proceedings:

SOCS-22 Volume 15

Volume

Issue:

Vol. 15 No. 1 (2022): Fifteenth International Symposium on Combinatorial Search

Track:

Extended Abstracts

Downloads:

Download PDF

Abstract:

This paper extends a study on improving the performance of reduction-based solvers for the problem of multi-agent pathfinding. The task is to navigate a set of agents in a graph without collisions. Solvers that reduce this problem to other formalisms often have issues scaling to larger instances in terms of the graph size. A previous study suggests that pruning the graph of most vertices based on a randomly chosen shortest path for each agent. In this paper, we study the effect of different choices of these paths.

DOI:

10.1609/socs.v15i1.21799


SOCS

Vol. 15 No. 1 (2022): Fifteenth International Symposium on Combinatorial Search



Published by , . All rights reserved.


Copyright ,