Importance sampling for reinforcement learning with multiple objectives
Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2001.
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | eng |
| Published: |
Massachusetts Institute of Technology
2014
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/1721.1/86774 |
| _version_ | 1826214156169117696 |
|---|---|
| author | Shelton, Christian R. (Christian Robert), 1975- |
| author2 | Tomaso Poggio. |
| author_facet | Tomaso Poggio. Shelton, Christian R. (Christian Robert), 1975- |
| author_sort | Shelton, Christian R. (Christian Robert), 1975- |
| collection | MIT |
| description | Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2001. |
| first_indexed | 2024-09-23T16:00:42Z |
| format | Thesis |
| id | mit-1721.1/86774 |
| institution | Massachusetts Institute of Technology |
| language | eng |
| last_indexed | 2024-09-23T16:00:42Z |
| publishDate | 2014 |
| publisher | Massachusetts Institute of Technology |
| record_format | dspace |
| spelling | mit-1721.1/867742019-04-12T09:05:35Z Importance sampling for reinforcement learning with multiple objectives Shelton, Christian R. (Christian Robert), 1975- Tomaso Poggio. Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science. Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science. Electrical Engineering and Computer Science. Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2001. Includes bibliographical references (p. 115-118). This thesis considers three complications that arise from applying reinforcement learning to a real-world application. In the process of using reinforcement learning to build an adaptive electronic market-maker, we find the sparsity of data, the partial observability of the domain, and the multiple objectives of the agent to cause serious problems for existing reinforcement learning algorithms. We employ importance sampling (likelihood ratios) to achieve good performance in partially observable Nlarkov decision processes with few data. Our importance sampling estimator requires no knowledge about the environment and places few restrictions on the method of collecting data. It can be used efficiently with reactive controllers, finite-state controllers, or policies with function approximation. We present theoretical analyses of the estimator and incorporate it into a reinforcement learning algorithm. Additionally, this method provides a complete return surface which can be used to balance multiple objectives dynamically. We demonstrate the need for multiple goals in a variety of applications and natural solutions based on our sampling method. The thesis concludes with example results from employing our algorithm to the domain of automated electronic market-making. by Christian Robert Shelton. Ph.D. 2014-05-07T17:05:00Z 2014-05-07T17:05:00Z 2001 2001 Thesis http://hdl.handle.net/1721.1/86774 49837623 eng M.I.T. theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission. See provided URL for inquiries about permission. http://dspace.mit.edu/handle/1721.1/7582 118 p. application/pdf Massachusetts Institute of Technology |
| spellingShingle | Electrical Engineering and Computer Science. Shelton, Christian R. (Christian Robert), 1975- Importance sampling for reinforcement learning with multiple objectives |
| title | Importance sampling for reinforcement learning with multiple objectives |
| title_full | Importance sampling for reinforcement learning with multiple objectives |
| title_fullStr | Importance sampling for reinforcement learning with multiple objectives |
| title_full_unstemmed | Importance sampling for reinforcement learning with multiple objectives |
| title_short | Importance sampling for reinforcement learning with multiple objectives |
| title_sort | importance sampling for reinforcement learning with multiple objectives |
| topic | Electrical Engineering and Computer Science. |
| url | http://hdl.handle.net/1721.1/86774 |
| work_keys_str_mv | AT sheltonchristianrchristianrobert1975 importancesamplingforreinforcementlearningwithmultipleobjectives |