Knowledge-based system developement for the game congkak / Syah Ali Reza Yaacob

Abstract

Game playing can be considered as a typical intelligent task, because games always form a closed environment with limited possibilities and clearly defined rules. This research studies a traditional game, Congkak. This game has 2 Rumah, 2 rows of 7 Kampung and each Kampung consist of 7 Buah. The objective is to implement Minimax and α-β pruning in this game. In this project, some of the game rules are modified in order to make it into a turn-based game. The original game requires both players to move simultaneously. Therefore, the Minimax algorithm is not applicable. The players are human and a goal-based agent. The agent will use Minimax and α-β pruning as the search algorithms. Game rules will act as the evaluation function, evaluating all moves created by the algorithms. The tree size is an approximate 3.6199 x 10¹⁸. As much as 100 test games are done within a week. The agent wins all game. This is due to logical errors in designing this game.

Metadata