Abstraction Hierarchies for Multi-Agent Pathfinding

<p>Multi-Agent Pathfinding is an NP-Complete search problem with a branching factor that is exponential in the number of agents. Because of this exponential feature, it can be difficult to solve optimally using traditional search techniques, even for relatively small problems. Many recent optimal solvers have attempted to reduce the complexity of the problem by resolving the conflicts between agent paths separately. Very little of this research has focused on creating quality heuristics to help solve the problem. In this thesis, we create heuristics using sub-problems created by removing agents from a complete problem instance. We combine this with the Independence Detection technique for solving the problem by separating agents into independent (non-conflicting) groups. The results showed moderate improvements in state expansions and computation time in problems with a large number of conflicting agents.</p>

Name thesis:degree_name

M.S.

Level thesis:degree_level

Masters Thesis

Year dc:date.available

2017

Author dc:creator

• Kraft, Aaron R.

Contributors dc:contributor

• Nathan Sturtevant, Ph.D.
• Mario Lopez
• Jun Zhang

• Heuristic
• Heuristic search
• MAPF
• Multi-agent pathfinding
• Pathfinding
• Path planning
• Computer Engineering

Statement dc:rights

• <p>Copyright is held by the author. User is responsible for all copyright compliance.</p>

Language dc:language

en

Repository record dc:identifier

https://digitalcommons.du.edu/etd/1247

OAI identifier oai:identifier

oai:digitalcommons.du.edu:etd-2247