Using Reinforcement Learning in Partial Order Plan Space

Access: Use of this item is restricted to the UNT Community
Description:

Partial order planning is an important approach that solves planning problems without completely specifying the orderings between the actions in the plan. This property provides greater flexibility in executing plans; hence making the partial order planners a preferred choice over other planning methodologies. However, in order to find partially ordered plans, partial order planners perform a search in plan space rather than in space of world states and an uninformed search in plan space leads to poor efficiency. In this thesis, I discuss applying a reinforcement learning method, called First-visit Monte Carlo method, to partial order planning in order to design agents which do not need any training data or heuristics but are still able to make informed decisions in plan space based on experience. Communicating effectively with the agent is crucial in reinforcement learning. I address how this task was accomplished in plan space and the results from an evaluation of a blocks world test bed.

Creator(s): Ceylan, Hakan
Creation Date: May 2006
Partner(s):
UNT Libraries
Collection(s):
UNT Theses and Dissertations
Usage:
Total Uses: 52
Past 30 days: 0
Yesterday: 0
Creator (Author):
Publisher Info:
Publisher Name: University of North Texas
Place of Publication: Denton, Texas
Date(s):
  • Creation: May 2006
  • Digitized: April 22, 2008
Description:

Partial order planning is an important approach that solves planning problems without completely specifying the orderings between the actions in the plan. This property provides greater flexibility in executing plans; hence making the partial order planners a preferred choice over other planning methodologies. However, in order to find partially ordered plans, partial order planners perform a search in plan space rather than in space of world states and an uninformed search in plan space leads to poor efficiency. In this thesis, I discuss applying a reinforcement learning method, called First-visit Monte Carlo method, to partial order planning in order to design agents which do not need any training data or heuristics but are still able to make informed decisions in plan space based on experience. Communicating effectively with the agent is crucial in reinforcement learning. I address how this task was accomplished in plan space and the results from an evaluation of a blocks world test bed.

Degree:
Level: Master's
Discipline: Computer Science
Language(s):
Subject(s):
Keyword(s): artificial intelligence | reinforcement learning | partial order planning
Contributor(s):
Partner:
UNT Libraries
Collection:
UNT Theses and Dissertations
Identifier:
  • OCLC: 70890296 |
  • ARK: ark:/67531/metadc5232
Resource Type: Thesis or Dissertation
Format: Text
Rights:
Access: Use restricted to UNT Community
License: Copyright
Holder: Ceylan, Hakan
Statement: Copyright is held by the author, unless otherwise noted. All rights reserved.