Multi-Agent Pathfinding: Definitions, Variants, and Benchmarks

Abstract

The multi-agent pathfinding problem (MAPF) is the fundamental problem of planning paths for multiple agents, where the key constraint is that the agents will be able to follow these paths concurrently without colliding with each other. Applications of MAPF include automated warehouses, autonomous vehicles, and robotics. Research on MAPF has been flourishing in the past couple of years. Different MAPF research papers assume different sets of assumptions, e.g., whether agents can traverse the same road at the same time, and have different objective functions, e.g., minimize makespan or sum of agents' actions costs. These assumptions and objectives are sometimes implicitly assumed or described informally. This makes it difficult for establishing appropriate baselines for comparison in research papers, as well as making it difficult for practitioners to find the papers relevant to their concrete application. This paper aims to fill this gap and facilitate future research and practitioners by providing a unifying terminology for describing the common MAPF assumptions and objectives. In addition, we also provide pointers to two MAPF benchmarks. In particular, we introduce a new grid-based benchmark for MAPF, and demonstrate experimentally that it poses a challenge to contemporary MAPF algorithms.

Effectively Incorporating Weighted Cost-to-go Heuristic in Suboptimal CBS (Extended Abstract)

“…We test our methods with different numbers of agents, in increments of 50, on 8 diverse maps from Stern et al (2019) and report mean values across 5 seeds. We use w so = 2 and a timeout of 300 seconds and report the speed up S method = T baseline /T method (larger is better) to normalize differences in hardware, where the baseline is EECBS.…”

Lower and Upper Bounds for Multi-Agent Multi-Item Pickup and Delivery: When a Decoupled Approach is Good Enough (Extended Abstract)

“…Each agent can carry multiple items at the same time, up to a given capacity parameter c >= 1. We use the flowtime, also known as sumof-costs, as the objective function (Stern et al 2019).…”

“…We modified both algorithms to be able to solve MAPF instances where each agent has to visit multiple intermediate goals before stopping at the final goal, instead of only requiring to move to its destination. To solve the MAPF instances, we used a classical MAPF problem statement (Stern et al 2019) when time is discretized as all the agents can perform only cardinal moves or wait in place. Furthermore, agents occupy their final goal after reaching it and do not disappear.…”

“…The TA can be formulated as the Vehicle Routing Problem (VRP) (Gendreau et al 2008) and the latter as the Multi-Agent Pathfinding (MAPF) (Stern et al 2019). Solving these two problems optimally is, however, NP-hard (Yu and LaValle 2016) and many solvers focus on quickly finding near-optimal solutions.…”

Quasi‐oppositional wild horse optimization based multi‐agent path finding scheme for real time IoT systems

“…Several researches have been conducted so far to manage the resources, particularly memory and energy (Yousefi et al, 2020). Researchers have proposed different approaches to achieve the optimal solution amongst which optimization algorithms (Stern et al, 2019) gained much attention in recent years. Metaheuristic is a common optimization algorithm in which the primary aim is to find a solution for the problem instead of the optimal path to reach the solution.…”