Engineering Commons^™

Abstraction Hierarchies For Multi-Agent Pathfinding, Aaron R. Kraft

Electronic Theses and Dissertations

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 …

Bidirectional Lao* Algorithm (A Faster Approach To Solve Goal-Directed Mdps), Venkata Deepti Kiran Bhuma

University of Kentucky Master's Theses

Uncertainty is a feature of many AI applications. While there are polynomial-time algorithms for planning in stochastic systems, planning is still slow, in part because most algorithms plan for all eventualities. Algorithms such as LAO* are able to find good or optimal policies more quickly when the starting state of the system is known.

In this thesis we present an extension to LAO*, called BLAO*. BLAO* is an extension of the LAO* algorithm to a bidirectional search. We show that BLAO* finds optimal or ^_E-optimal solutions for goal-directed MDPs without necessarily evaluating the entire state space. BLAO* …

Solution To A Multicriteria Aircraft Routing Problem Utilizing Parallel Search Techniques, James J. Grimm Iii

Theses and Dissertations

Pilots select routes based on factors such as threats, fuel, time on target, distance, and refueling points. This is a time consuming task. This thesis presents the software engineering synthesis of a software tool, based on a parallelized A* search algorithm, to select routes. For simplicity only threats and distance are used. A centralized open list is used with one processor managing the list while the other processors perform the node expansions. This decomposition result in a dynamically load balanced system. A number of parameters are changed to study their impact on the execution time. The use of a branch …