Robust Multi-Agent Path Finding and Executing

Abstract

Multi-agent path-finding (MAPF) is the problem of finding a plan for moving a set of agents from their initial locations to their goals without collisions. Following this plan, however, may not be possible due to unexpected events that delay some of the agents. In this work, we propose a holistic solution for MAPF that is robust to such unexpected delays. First, we introduce the notion of a k-robust MAPF plan, which is a plan that can be executed even if a limited number (k) of delays occur. We propose sufficient and required conditions for finding a k-robust plan, and show how to convert several MAPF solvers to find such plans. Then, we propose several robust execution policies. An execution policy is a policy for agents executing a MAPF plan. An execution policy is robust if following it guarantees that the agents reach their goals even if they encounter unexpected delays. Several classes of such robust execution policies are proposed and evaluated experimentally. Finally, we present robust execution policies for cases where communication between the agents may also be delayed. We performed an extensive experimental evaluation in which we compared different algorithms for finding robust MAPF plans, compared different ro- bust execution policies, and studied the interplay between having a robust plan and the performance when using a robust execution policy.

Multi-Agent Path Finding for Large Agents

“…We show in Section 3.2 that this kind of direct adaptation is inefficient in many cases. Using a different perspective, Atzmon et al (2018) generalized the definition of a conflict from a timestep to a time range in order to obtain robust plans.…”

“…Atzmon et al (2018) introduced similar asymmetric and symmetric approaches to add multiple constraints, although they studied a different problem and added constraints for the same vertex at different timesteps.…”

Multi-agent path finding on real robots

“…To minimize these effects, we may require the abstract plan to leave some space between the moving agents. We will use the notation from k-robust MAPF [1]. The k-robust plan is a valid MAPF plan that in addition requires for each vertex of the graph to be unoccupied for at least k time steps before another agent can enter it.…”

“…Specifically, we explore the very classical MAPF setting as described above, the k-robust setting [1], where a gap is required between the robots to compensate possible delays during execution, and finally a model that directly encodes turning operations (the classical setting does not assume direction of movement). The abstract plans are then translated to motion primitives, which consist of forward movement, turning left/right, and waiting.…”

Prioritized Multi-agent Path Finding for Differential Drive Robots

“…This is a preprint of the paper accepted to ECMR 2019: https://ieeexplore.ieee.org/document/8870957 algorithm itself in the way that it lifts as much constraints as possible [13]- [17] and/or produces robust solutions that are more likely to be executed safely [18]- [20].…”