About reversi-mcts

reversi-mcts is a reversi/otherllo agent that is based on the Monte Carlo Tree Search (MCTS) algorithm implemented in the Go programming language.

It includes both an API client and a javascript based interface to play against the agent.

Give it a go and test your skills against this agent at royhung.com/reversi

Installing reversi-mcts

Make sure you have a working Go installation, see the installation guide at http://golang.org/doc/install.html

Next, make sure $GOPATH is set (e.g. as ~/go) and:

$ cd $GOPATH/src
$ git clone https://gitlab.com/royhung_/reversi-monte-carlo-tree-search.git
$ cd reversi-monte-carlo-tree-search
$ go build main.go reversi.go api.go

Using reversi-mcts

Run the built package to start the server:

$ ./reversi-monte-carlo-tree-search

The application will be running on http://localhost:8080 with a reversi board interface and a playable MCTS agent.

API Endpoint

To access the agent's API directly, make a POST request to /search_move.
The endpoint consumes data on the state of the game via the positions of black pieces and white pieces in play.
The response from the MCTS agent will contain the coordinates of the position the agent will place its next move.

POST /search_move

Request POST JSON Example

{
    "blackFilled":[[3,3],[4,4]],    
    "whiteFilled":[[3,4],[4,3]],    
    "turn":1,                      
}

Property	Type	Description
blackFilled	Object	Array of coordinate positions [ i , j ] of black pieces, where i refers to the ith row on board and j refers to the jth row on the board
whiteFilled	Object	Array of coordinate positions [ i , j ] of white pieces, where i refers to the ith row on board and j refers to the jth row on the board
turn	Integer	The colour agent is supposed to play as for its turn (1 black, -1 white)

Response POST JSON Example

{
    "move":[3,2],                   
    "turn":1,                       
    "blackScore":1,
    "whiteScore":4
}

Property	Type	Description
move	Object	Array containing coordinate position [ i, j ] of the move the agent has made
turn	Integer	The colour of the piece placed (1 black, -1 white)
blackScore	Integer	The resulting number of black pieces on the board after move is made
whiteScore	Integer	The resulting number of white pieces on the board after move is made

More information

For a more detailed write up on the algorithm, performance and the parameters used. Please visit https://royhung.com/reversi