Author |
Topic: A Survey of Monte Carlo Tree Search Methods (Read 3465 times) |
|
Janzert
Forum Guru
Arimaa player #247
Gender: 
Posts: 1016
|
 |
A Survey of Monte Carlo Tree Search Methods
« on: Mar 2nd, 2012, 8:12am » |
 Quote  Modify
|
Given the interest in MCTS by members of the community I thought I'd pass this on here.
This is a new article out giving a broad overview of various MCTS techniques currently in use.
http://www.cameronius.com/cv/mcts-survey-master.pdf
|
|
 IP Logged |
|
|
|
Hippo
Forum Guru
Arimaa player #4450
Gender:
Posts: 883
|
|
Re: A Survey of Monte Carlo Tree Search Methods
« Reply #1 on: Mar 3rd, 2012, 12:53pm » |
Quote Modify
|
Wow, that is so long ... I have felt asleep several times while reading ... anyways thanks for the link.
It's surely good source of inspiration.
BTW: Kozelek's work is mentioned there  .
|
| « Last Edit: Mar 3rd, 2012, 2:55pm by Hippo » |
IP Logged |
|
|
|
tize
Forum Guru
Arimaa player #3121
Gender:
Posts: 118
|
|
Re: A Survey of Monte Carlo Tree Search Methods
« Reply #2 on: Mar 4th, 2012, 4:52am » |
Quote Modify
|
on Mar 3rd, 2012, 12:53pm, Hippo wrote:| It's surely good source of inspiration. |
|
I totally agree. I can't stop thinking of different ideas to try to merge alphabeta to mcts and/or the other way around. I really hope that all this will inspire someone to try (and succeed) with something new.
|
|
IP Logged |
|
|
|
Hippo
Forum Guru
Arimaa player #4450
Gender:
Posts: 883
|
|
Re: A Survey of Monte Carlo Tree Search Methods
« Reply #3 on: Mar 4th, 2012, 2:02pm » |
Quote Modify
|
Finally I have skipped the games section and finished the reading. Seems most of the ideas I have in mind were already studied in some context.
This would be rather easy programming, but very difficult optimisation in highly dimensional parameter space.
|
|
IP Logged |
|
|
|
|
Home
Help
Search
Members
Login
Register
Reply
Notify of replies
Send Topic
Print