In a multi-agent pathfinding (MAPF) problem, agents need to navigate from their start to their goal locations without colliding into each other. There are various MAPF algorithms, including Windowed Hierarchical Cooperative A*, Flow Annotated Replanning, and Bounded Multi-Agent A*. It is often the case that there is no a single algorithm that dominates all MAPF instances. Therefore, in this paper, we investigate the use of deep learning to automatically select the best MAPF algorithm from a portfolio of algorithms for a given MAPF problem instance. Empirical results show that our automatic algorithm selection approach, which uses an off-the-shelf convolutional neural network, is able to outperform any individual MAPF algorithm in our portfolio.
@article{arxiv.1906.03992,
title = {Automatic Algorithm Selection In Multi-agent Pathfinding},
author = {Devon Sigurdson and Vadim Bulitko and Sven Koenig and Carlos Hernandez and William Yeoh},
journal= {arXiv preprint arXiv:1906.03992},
year = {2019}
}