Date of Award

1-22-2020

Publication Type

Master Thesis

Degree Name

M.Sc.

Department

Computer Science

Supervisor

Scott Goodwin

Rights

info:eu-repo/semantics/openAccess

Abstract

Pathfinding is an essential part of navigation systems, often used in video games, route planning and robotic navigation. A* search has been one of the most well-known and frequently used algorithms for pathfinding. A* uses an open list and a closed list to keep track of all nodes generated and expanded. The size and performance of these data structures are major drawbacks of A*. Lookahead is used to investigate future outcomes and improve the quality of available choices. Lookaheads are done on a DFS manner from the frontier of A* search. This combination of A* and DFS lookahead has been shown to save space when working with puzzles. We leverage this concept with grid-based pathfinding in video games to save the amount of space consumed. However, because grids contain redundant paths, the DFS lookaheads end up being an overhead as they do not maintain a list of nodes visited or expanded. By using a domain-specific pruning technique, we significantly improve the time taken by the algorithm and further improve upon the space consumed. A combination of lookahead and A* search with this pruning technique is, therefore, able to achieve improvement in both space-consumed and time-taken over the standard A* search algorithm for grid-based pathfinding.

Share

COinS