Push and rotate: cooperative multi-agent path planning

Authors:
Boris de Wilde;Adriaan W. ter Mors;Cees Witteveen
Affiliations:
Delft University of Technology, Delft, Netherlands;Delft University of Technology, Delft, Netherlands;Delft University of Technology, Delft, Netherlands
Venue:
Proceedings of the 2013 international conference on Autonomous agents and multi-agent systems
Year:
2013

Citing 9
Cited 0

Toward efficient trajectory planning: the path-velocity decomposition

International Journal of Robotics Research
Algorithm 447: efficient algorithms for graph manipulation

Communications of the ACM
High quality navigation in computer games

Science of Computer Programming
Coordinating Pebble Motion On Graphs, The Diameter Of Permutation Groups, And Applications

SFCS '84 Proceedings of the 25th Annual Symposium onFoundations of Computer Science, 1984
Exploiting subgraph structure in multi-robot path planning

Journal of Artificial Intelligence Research
A novel approach to path planning for multiple robots in bi-connected graphs

ICRA'09 Proceedings of the 2009 IEEE international conference on Robotics and Automation
Finding the shortest move-sequence in the graph-generalized 15-puzzle is NP-hard

Studies in complexity and cryptography
MAPP: a scalable multi-agent path planning algorithm with tractability and completeness guarantees

Journal of Artificial Intelligence Research
Complete algorithms for cooperative pathfinding problems

IJCAI'11 Proceedings of the Twenty-Second international joint conference on Artificial Intelligence - Volume Volume One

Quantified Score

Hi-index	0.00

Visualization

Abstract

In cooperative multi-agent path planning, agents must move between start and destination locations and avoid collisions with each other. Many recent algorithms require some sort of restriction in order to be complete, except for the Push and Swap algorithm [Luna and Bekris, 2011], which claims only to require two unoccupied locations in a connected graph. Our analysis shows, however, that for certain types of instances Push and Swap may fail to find a solution. We present the Push and Rotate algorithm, an adaptation of the Push and Swap algorithm, and prove that by fixing the latter's shortcomings, we obtain an algorithm that is complete for the class of instances with two unoccupied locations in a connected graph. In addition, we provide experimental results that show our algorithm to perform competitively on a set of benchmark problems from the video game industry.