Multi-Agent Path Finding for Non-Conflicting Paths: An Infinite Speed Conflict-Based Search Approach

Yaodong Hou, Gang Tao, Zheng Zang, Guojie Kong, Jianwei Gong*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Multi-Agent Path Finding (MAPF) is the problem of finding a set of non-conflicting paths for multiple agents on a graph, which is NP-hard to solve optimally. This paper proposes an infinite speed Conflict-Based Search (IS-CBS) algorithm to solve the MAPF problem when the agents move at different speeds and the edges correspond to roads of varying lengths. In addition, three distinct approaches are introduced to calculate admissible heuristics for IS-CBS: the sum approach which uses the sum of all heuristics as the final heuristic, the Pareto approach which builds a Pareto set and selects nodes from it, and the alternation approach which alternates between heuristics in search iterations. In our experiment, IS-CBS with heuristics outperforms IS-CBS in terms of success rate, runtime, and number of expanded nodes which is reduced by up to a factor of 21.

Original languageEnglish
Title of host publication2023 42nd Chinese Control Conference, CCC 2023
PublisherIEEE Computer Society
Pages5500-5505
Number of pages6
ISBN (Electronic)9789887581543
DOIs
Publication statusPublished - 2023
Event42nd Chinese Control Conference, CCC 2023 - Tianjin, China
Duration: 24 Jul 202326 Jul 2023

Publication series

NameChinese Control Conference, CCC
Volume2023-July
ISSN (Print)1934-1768
ISSN (Electronic)2161-2927

Conference

Conference42nd Chinese Control Conference, CCC 2023
Country/TerritoryChina
CityTianjin
Period24/07/2326/07/23

Keywords

  • Multi-Agent Path Finding
  • heuristic search
  • infinite speed Conflict-Based Search

Fingerprint

Dive into the research topics of 'Multi-Agent Path Finding for Non-Conflicting Paths: An Infinite Speed Conflict-Based Search Approach'. Together they form a unique fingerprint.

Cite this