A Distributed AI Aided 3D Domino Game

TL;DR

This paper presents a distributed 3D domino game played over a network with four players, including four levels of AI with increasing complexity, from basic rule-based to self-developing master AI.

Contribution

It introduces a novel distributed network implementation of a 3D domino game with multiple AI levels, including self-developing AI, using Microsoft XNA and TCP/IP.

Findings

Four AI levels developed, from child-like to master AI.

Distributed system enables multiplayer gameplay over network.

AI adapts to opponents' skill levels for improved performance.

Abstract

In the article a turn-based game played on four computers connected via network is investigated. There are three computers with natural intelligence and one with artificial intelligence. Game table is seen by each player's own view point in all players' monitors. Domino pieces are three dimensional. For distributed systems TCP/IP protocol is used. In order to get 3D image, Microsoft XNA technology is applied. Domino 101 game is nondeterministic game that is result of the game depends on the initial random distribution of the pieces. Number of the distributions is equal to the multiplication of following combinations: . Moreover, in this game that is played by four people, players are divided into 2 pairs. Accordingly, we cannot predict how the player uses the dominoes that is according to the dominoes of his/her partner or according to his/her own dominoes. The fact that the natural intelligence can be a player in any level affects the outcome. These reasons make it difficult to develop an AI. In the article four levels of AI are developed. The AI in the first level is equivalent to the intelligence of a child who knows the rules of the game and recognizes the numbers. The AI in this level plays if it has any domino, suitable to play or says pass. In most of the games which can be played on the internet, the AI does the same. But the AI in the last level is a master player, and it can develop itself according to its competitors' levels.