Improving the Sum-of-Cost Methods for Reduction-Based Multi-Agent Pathfinding Solvers

Roland Kaminski, Torsten Schaub, Klaus Strauch, Jiří Švancara

2024

Abstract

Multi-agent pathfinding is the task of guiding a group of agents through a shared environment while preventing collisions. This problem is highly relevant in various real-life scenarios, such as warehousing, robotics, navigation, and computer games. Depending on the context in which the problem is applied, we may have specific criteria for the quality of a solution, expressed as a cost function. The most common cost functions are the makespan and sum-of-cost. Minimizing either of them is computationally challenging, leading to the development of numerous approaches for solving multi-agent pathfinding. In this paper, we explore reduction-based solving under the sum-of-cost objective. We introduce a reduction to answer set programming (ASP) using two existing approaches for sum-of-cost minimization, originally introduced for a reduction to Boolean satisfiability (SAT). We propose several enhancements and use the Clingo ASP system to implement them. Experiments show that these enhancements significantly improve performance. Particularly, the performance on larger maps increases in comparison to the original variants.

Paper Citation

in Harvard Style

Kaminski R., Schaub T., Strauch K. and Švancara J. (2024). Improving the Sum-of-Cost Methods for Reduction-Based Multi-Agent Pathfinding Solvers. In Proceedings of the 16th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART; ISBN 978-989-758-680-4, SciTePress, pages 264-271. DOI: 10.5220/0012353500003636

in Bibtex Style

@conference{icaart24,
author={Roland Kaminski and Torsten Schaub and Klaus Strauch and Jiří Švancara},
title={Improving the Sum-of-Cost Methods for Reduction-Based Multi-Agent Pathfinding Solvers},
booktitle={Proceedings of the 16th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART},
year={2024},
pages={264-271},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012353500003636},
isbn={978-989-758-680-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 16th International Conference on Agents and Artificial Intelligence - Volume 1: ICAART
TI - Improving the Sum-of-Cost Methods for Reduction-Based Multi-Agent Pathfinding Solvers
SN - 978-989-758-680-4
AU - Kaminski R.
AU - Schaub T.
AU - Strauch K.
AU - Švancara J.
PY - 2024
SP - 264
EP - 271
DO - 10.5220/0012353500003636
PB - SciTePress