Minimax implementation (mgechev#171)

* minimax

* remove blank line

Co-Authored-By: Minko Gechev <mgechev@gmail.com>

* cr fixes

* simple game tests added

Co-authored-by: Minko Gechev <mgechev@gmail.com>

Author: krzysztof-grzybek

src/others/minimax.js

@@ -0,0 +1,120 @@
+(function (exports) {
+  'use strict';
+  /* eslint max-params: 0 */
+
+  /**
+  * @param {Function} getPossibleNextStatesFn Function which returns all possible next moves with states .
+  * @param {Function} isGameOverFn Function which returns if game is over.
+  * @param {Function} getScoreFn Function which returns score.
+  * @return {Function} minimax function
+   */
+  function minimaxBuilder(
+    getPossibleNextStatesFn,
+    isGameOverFn,
+    getScoreFn
+  ) {
+    /**
+     * Minimax (sometimes MinMax, MM[1] or saddle point[2]) is a decision rule used in artificial intelligence,
+     * decision theory, game theory, statistics, and philosophy for minimizing the possible loss for a worst case (maximum loss) scenario.
+     * Optimized with alpha-beta pruning.
+     *  {@link https://en.wikipedia.org/wiki/Minimax}
+     *  {@link https://en.wikipedia.org/wiki/Alpha%E2%80%93beta_pruning}
+     *
+     * @public
+     * @module others/minimax
+     *
+     * @example
+     *
+     * var miniMax =
+     *  require('path-to-algorithms/src/others/minimax').minimax;
+     * var result = minimax(
+     *   [1, 2, 3],
+     *   true,
+     *   5,
+     *   -Infinity,
+     *   Infinity,
+     *   state => ({ move: 0, state: [2, 3, 4] }),
+     *   state => state[1] < 3,
+     *   state => state[1]
+     * );
+     *
+     * @param {*} state Current game state
+     * @param {Boolean} maximize Defines if the result should be maximized or minimized
+     * @param {Number} depth Defines the maximum depth search
+     * @param {Number} alpha Maximum score that the minimizing player is assured
+     * @param {Number} beta Minimum score that the maximizing player is assured
+     * @return {{score: Number, move: *}} which contains the minimum coins from the given
+     *  list, required for the change.
+     */
+    const minimax = (
+      state,
+      maximize,
+      depth,
+      alpha,
+      beta
+    ) => {
+      if (depth === 0 || isGameOverFn(state)) {
+        const score = getScoreFn(state);
+        return {score, move: null};
+      }
+
+      const possibleMoveResults = getPossibleNextStatesFn(state);
+
+      if (maximize) {
+
+        let maxResult = {score: -Infinity, move: null};
+
+        for (const next of possibleMoveResults) {
+          const result = minimax(
+            next.state,
+            false,
+            depth - 1,
+            alpha,
+            beta,
+          );
+
+          if (result.score > maxResult.score) {
+            maxResult = {score: result.score, move: next.move};
+          }
+
+          alpha = Math.max(alpha, result.score);
+
+          if (alpha >= beta) {
+            break;
+          }
+        }
+
+        return maxResult;
+      } else {
+        let minResult = {score: Infinity, move: null};
+
+        for (const next of possibleMoveResults) {
+          const result = minimax(
+            next.state,
+            true,
+            depth - 1,
+            alpha,
+            beta,
+          );
+
+          if (result.score < minResult.score) {
+            minResult = {score: result.score, move: next.move};
+          }
+
+          beta = Math.min(beta, result.score);
+
+          if (beta <= alpha) {
+            break;
+          }
+        }
+
+        return minResult;
+      }
+    }
+
+    return minimax;
+  }
+
+  exports.minimaxBuilder = minimaxBuilder;
+
+})(typeof window === 'undefined' ? module.exports : window);

test/data-structures/avl-tree.spec.js

@@ -178,8 +178,8 @@ describe('AVL Tree', function () {
     avlTree.insert(6);
     avlTree.remove(32);
     avlTree.remove(11);
-    avlTree.remove(25);    
-    
+    avlTree.remove(25);
+
     // depth 1
     expect(avlTree._root.value).toBe(37);
     expect(avlTree._root._height).toBe(4);