ChessCorp Artificial Intelligence: Kong

Abstract

This project is an example of an Artificial Intelligence module.

It uses the following modules from ChessCorp project:

• Chess Rules set View

Engine

The AI search engine is based on the [Alpha Beta Pruning algorithm] (https://en.wikipedia.org/wiki/Alpha%E2%80%93beta_pruning). It chooses best move by simulating in-depth plies and refuting moves that cannot alter its final decision.

Strategies

The AI implements several strategies that can be configured as an option. Strategies are based on material and positions. They influence the board evaluation and as a matter of fact the AI's way of playing.

Basic Strategy

The basic strategy is based on the common principles about chess programming.

-Pawns progress forward. -Bishops and Knights control center, avoid corners and extreme corners. -Queen and Rook avoid center. -King hide in the back.

Random Strategy

The random strategy does not take into account material and position. Losing a pawn is as bad as losing a queen.

Configuration

The AI can be configured be setting options:

• depth: The depth of the search algorithm. A higher depth provide a deeper and longer thinking from the AI (Integer, 3 by default).
• monitor: Enable/disable the monitoring. Monitoring might be time-consuming because it generates PGN test and stores searched nodes. (Boolean, false by default).
• strategy: Set the strategy to use to evaluate board and moves (String, 'basic' by default).
• timeout: Set up a timeout after which the AI shall return a move (Integer, 10000 by default).

Install

As a Node module

$ npm install chess-ai-kong

This will install chess-ai-kong module in the node_modules folder.

Usage

Import the main ai-kong object

> var chessAI = require('chess-ai-kong');

> chessAI
{ 
  playPosition: [Function: playPosition],
  play: [Function: play],
  setOptions: [Function: setOptions]
}

Configure

 > chessAI.setOptions(
{
  depth: 4,
  monitor: true,
  strategy: 'basic',
  timeout: 10000
}
);

Play after a sequence of moves

The AI starts from an initial position and replays every moves provided before searching for the next move. The method returns a PGN move. PGN move text is described here https://en.wikipedia.org/wiki/Portable_Game_Notation

> var move = chessAi.play(['e4']);
'Nf6'

Note: The AI search a move for the player's turn after the sequence of moves. If there is an odd number of moves, the AI searches for a WHITE move, otherwise, it searches for a BLACK move.

Play after a certain position

The AI starts from the given position and search for the next move. The method returns a PGN move. PGN move text is described here https://en.wikipedia.org/wiki/Portable_Game_Notation

> var move = chessAi.playPosition(position);
'e4'

Note: The AI search a move for the player's turn. For example, if the initial position is passed in parameter, the AI will search for a WHITE move.