Push and rotate: a complete multi-agent pathfinding algorithm

作者: B. de Wilde , A. W. Ter Mors , C. Witteveen

DOI: 10.1613/JAIR.4447

关键词:

摘要: Multi-agent Pathfinding is a relevant problem in wide range of domains, for example robotics and video games research. Formally, the considers graph consisting vertices edges, set agents occupying vertices. An agent can only move to an unoccupied, neighbouring vertex, finding minimal sequence moves transfer each from its start location destination NP-hard problem. We present Push Rotate, new algorithm that complete problems which there are at least two empty Rotate first divides into subgraphs within it possible reach any position subgraph, then uses simple push, swap, rotate operations find solution; post-processing also presented eliminates redundant moves. be seen as extending Luna Bekris's Swap algorithm, we showed incomplete previous publication. In our experiments compare approach with Swap, MAPP, Bibox algorithms. The latter restricted smaller class instances requires biconnected graphs, but nevertheless considered state art due strong performance. Our show suffers incompleteness, MAPP generally not competitive better than on randomly generated instances, while performs grids.

参考文章(45)
A. W. ter Mors, C. Witteveen, J. Zutt, Context-aware logistic routing and scheduling international conference on automated planning and scheduling. pp. 328- 335 ,(2007)
Ryan Luna, Kostas E. Bekris, Push and swap: fast cooperative path-finding with completeness guarantees international joint conference on artificial intelligence. pp. 294- 300 ,(2011) , 10.5591/978-1-57735-516-8/IJCAI11-059
B. De Wilde, Cooperative Multi-Agent Path Planning TU Delft, Delft University of Technology. ,(2012)
David Šišlák, Michal Pěchouček, Přemysl Volf, Dušan Pavlíček, Jiří Samek, Vladimír Mařík, Paul Losiewicz, AGENTFLY: Towards Multi-Agent Technology in Free Flight Air Traffic Control Birkhäuser Basel. pp. 73- 96 ,(2007) , 10.1007/978-3-7643-8571-2_5
Cees Witteveen, Adriaan W. ter Mors, Boris de Wilde, Push and rotate: cooperative multi-agent path planning adaptive agents and multi-agents systems. pp. 87- 94 ,(2013) , 10.5555/2484920.2484938
Oded Goldreich, Finding the shortest move-sequence in the graph-generalized 15-puzzle is NP-hard Studies in complexity and cryptography. pp. 1- 5 ,(2011) , 10.1007/978-3-642-22670-0_1
Sebastian Trüg, Jörg Hoffmann, Bernhard Nebel, Applying Automatic Planning Systems to Airport Ground-Traffic Control – A Feasibility Study Lecture Notes in Computer Science. pp. 183- 197 ,(2004) , 10.1007/978-3-540-30221-6_15
Trevor Standley, Finding optimal solutions to cooperative pathfinding problems national conference on artificial intelligence. pp. 173- 178 ,(2010)
Adriaan W. ter Mors, Cees Witteveen, Jonne Zutt, Fernando A. Kuipers, Context-aware route planning multiagent system technologies. ,vol. 6251, pp. 138- 149 ,(2010) , 10.1007/978-3-642-16178-0_14