A Study of Solving Life and Death Problem for 7x7 Kill-All Go¬¬
碩士 === 國立交通大學 === 多媒體工程研究所 === 101 === Go is one of the most complicated two-player zero-sum perfect information games. The strength of computer Go get to the level of amateur 3or 4 dan since people apply MCTS to the computer Go in 2006. Life and death problem is very important to Go. The compu...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2013
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/41796122985626839062 |
| Summary: | 碩士 === 國立交通大學 === 多媒體工程研究所 === 101 === Go is one of the most complicated two-player zero-sum perfect information games. The strength of computer Go get to the level of amateur 3or 4 dan since people apply MCTS to the computer Go in 2006.
Life and death problem is very important to Go. The computer Go will be stronger if we can compute life and death problem more accurately.
In this paper, we implement one MCTS Solver specialized in solving life and death problem, and illustrate how to combine MCTS Solver with threat space search and lambda search.
|
|---|