Date of Award

5-1-2012

Degree Name

Master of Science

Department

Computer Science

First Advisor

Hexmoor, Henry

Abstract

Learning by dynamic programming, greedy algorithm, and graph theory requires prior information like distance and elapse time, which are both costs, to produce optimal path (i.e., the solution). However, there is no prior information (distance and/or elapsed time) in most situations until the environment is fully explored. There is a need for the system to learn from the agent's (i.e., human's) most recent decisions and reproduce the best path (i.e., solution) based on the agent's past experiences and the current situation. The purpose of this thesis is to propose a solution to improve path planning by observing human decisions in path planning and chunking pieces that are superior to stored paths. This includes cumulatively storing the best path chunks based on the agent's experiences.

Share

COinS
 

Access

This thesis is only available for download to the SIUC community. Others should
contact the interlibrary loan department of your local library.