Recent Changes - Search:


Home Page
MAPF Info
MAPF News
Mailing List
Meetings
Publications
Researchers
Benchmarks
Software
Apps
Tutorials
Class Projects

[Internal]

Publication

E. Lam, P. Stuckey and D. Harabor. Optimal Multi-Agent Pickup and Delivery Using Branch-and-Cut-and-Price Algorithms. Transportation Science, (in print), 2024.


Abstract: Given a set of agents and a set of pickup-delivery requests located on a two-dimensional grid map, the multi-agent pickup and delivery problem assigns the requests to the agents such that every agent moves from its start location to the locations of its assigned requests and finally, to its end location without colliding into other agents and that the sum of arrival times is minimized. This paper proposes two exact branch-and-cut-and-price algorithms for the problem. The first algorithm performs a three-level search. A high-level master problem selects an optimal sequence of requests and a path for every agent from a large collection. A mid-level sequencing problem and a low-level navigation problem are solved simultaneously to incrementally enlarge the collection of request sequences and paths. The second algorithm first solves the sequencing problem to find a set of request sequences and then solves the navigation problem to determine if paths compatible with the request sequences exist. Experimental results indicate that the integrated algorithm solves more instances with higher congestion, and the deferred algorithm solves more instances with lower congestion and could scale to 100 agents and 100 requests, significantly higher than a state-of-the-art suboptimal approach.


Download the paper in pdf.

Edit - History - Print - Recent Changes - Search
Page last modified on February 16, 2025, at 10:56 AM