Year: 2021
Type: conference-paper
Abstract: In a multi-agent path-finding (MAPF) problem, the task is to find 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. Guaranteeing that collisions will never occur may be impossible. An important task is to find a plan that is very likely to succeed, even though unexpected delays may occur. We propose an algorithm for finding a plan in which the probability that no collisions will occur is at least a given parameter p (p-robust plan). We show that finding an optimal p-robust plan is significantly more difficult than finding an optimal standard plan. As a practical solution, we propose a greedy algorithm based on the Conflict-Based Search framework. Our experiments show that it finds p-robust plans with cost that is relatively close to the optimal cost of the standard, non-robust plans.
Authors: Dor Atzmon, Ariel Felner, Roni Stern
Institution: Ben-Gurion University of the Negev
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
Funder: National Science Foundation
Awards: CPS: Small: Novel Algorithmic Techniques for Drone Flight Planning on a Large Scale, NSF-BSF:RI:Small:Collaborative Research:Next-Generation Multi-Agent Path Finding Algorithms, S&AS: FND: Long-Term Planning and Robust Plan Execution for Multi-Robot Systems, Development of a Modular, Scalable, and Extensible Model-Based Systems Engineering Advanced Manufacturing Curriculum, NSF-BSF:RI:Small:Collaborative Research:Next-Generation Multi-Agent Path Finding Algorithms, RI: Medium: Collaborative Research: Experience-Based Planning: A Framework for Lifelong Planning
OpenAlex ID: W2965068353
