negamax-alpha-beta

A fast, modular JavaScript implementation of the negamax depth-first tree search algorithm with alpha-beta pruning.

About

This module searches the game tree of any zero-sum two-player game in an efficient manner, finding the optimal move for the current player-to-move in the given gameState. You can model your gameState and move objects however you see fit. They can be javascript objects, numbers, strings, anything you want. The requirements of the functions you must supply are detailed below in the Usage section.

Note: this implementation assumes in-place modification of the game states in the user-supplied makeMove and unmakeMove functions. This tends to be faster and less memory-intensive than having a single makeMove function that clones the game state and returns a new, modified game state. It is critical that the makeMove and unmakeMove match one another perfectly. Calling makeMove(gameState, move) and then unmakeMove(gameState, move) must result in an identical gameState to the one before calling the two functions.

Installation

npm install negamax-alpha-beta

Usage

const NegamaxAlphaBeta = require("negamax-alpha-beta");

let config = {
  generateMoves: (gameState) => { return [] },
  makeMove: (gameState, move) => { return true },
  unmakeMove: (gameState, move) => { return },
  evaluate: (gameState) => { return 0 },
  evaluateTerminal: (gameState) => { return null }
};

let negamax = new NegamaxAlphaBeta(config);

let result = negamax.search(gameState, depth);

console.log(`Result: score = ${result.score}, bestMove = ${result.bestMove}`);

Constructor: NegamaxAlphaBeta (config) { }

The constructor takes one argument, a configuration object containing several user-defined functions that operate on your generic gameState object:

generateMoves: function (gameState) { }

Your generateMoves function must take a gameState object and return an array of all legal moves for that game state. You can represent the moves however you see fit: integers, objects, strings, etc.

makeMove: function (gameState, move) { }

Your makeMove function must take a gameState object and a move object, perform the move upon the gameState, altering it in place, and return a boolean value that represents whether or not the side-to-move has changed after having performed the move.

unmakeMove: function (gameState, move) { }

Your unmakeMove function must take a gameState object and a move object, un-perform the move upon the gameState, altering it in place. This must end up producing the exact same gameState as before having called makeMove.

evaluate: function (gameState) { }

Your evaluate function must take a gameState object and return a numeric value representing the score of the gameState from the perspective of the gameState's current player-to-move. Higher numbers mean the gameState is better for the current player-to-move.

evaluateTerminal: function (gameState) { }

Your evaluateTerminal function must take a gameState object and either return null if the gameState is not in a terminal state (i.e. the game is not over), or return a numeric value representing the the score of the terminal gameState from the perspective of the gameState's current player-to-move. Higher numbers mean the gameState is better for the current player-to-move. Typically you would return 0 for a draw, some extreme positive number for a win, or some extreme negative number for a loss. The values returned for win and loss should be more extreme than any possible return value from calling evaluate on non-terminal states.

Function: search (gameState, depth) { }

The search method takes a gameState object and a depth integer, searches the game tree and returns an object like { score: 1.2345, bestMove: {some move object} }.

License

The MIT License (MIT)

Copyright (c) 2019 Aaron Hanson

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.