Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/B28C3C09-4B40-460B-8E86-BEC27920D173 | ||
| Year | 2019 | ||
| Type of Item | Master Thesis | ||
| License |
|
||
| Bibliographic Citation | Ioannis Skoulakis, "Efficient reinforcement learning in adversarial games", Master Thesis, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2019 https://doi.org/10.26233/heallink.tuc.83853 | ||
| Appears in Collections |
The ability of learning is critical for agents designed to compete in a variety of two-player, turn-taking, tactical adversarial games, such as Backgammon, Othello/Reversi, Chess, Hex, etc. The mainstream approach to learning in such games consists of updating some state evaluation function usually in a Temporal Difference (TD) sense either under the MiniMax optimality criterion or under optimization against a specific opponent. However, this approach is limited by several factors: (a) updates to the evaluation function are incremental, (b) stored samples from past games cannot be utilized, and (c) the quality of each update depends on the current evaluation function due to bootstrapping. In this thesis, we present four variations of a learning approach based on the Least-Squares Policy Iteration (LSPI) algorithm that overcome these limitations by focusing on learning a state-action evaluation function. The key advantage of the proposed approaches is that the agent can make batch updates to the evaluation function with any collection of samples, can utilize samples from past games, and can make updates that do not depend on the current evaluation function since there is no bootstrapping. We demonstrate the efficiency and the competency of the LSPI agents over the TD agent and selected benchmark opponents in the classical board games of Othello/Reversi and Backgammon.
Copyright © DIAS 2013
