Adaptive AI Pada Survival Horror Game Menggunakan Fuzzy

Abstract

There are many methods used by developers to create artificial intelligence in their games. The method most often used by developers because of it’s flexibility is the scripting method. Scripting method is done by writing several rules in the source code in the form of if - then. The problem with this method is that the NPC’s artificial intelligence is easily exploited by gamer who have understood the pattern of the NPC’s behavior.

Therefore the fuzzy logic is used to overcome these problems.The use of the fuzzy inference method is to make the AI to be adaptive to certain situation where depending on the situation the AI will give different decision making. In this case, Tsukamoto Fuzzy inference method is used in this design. In the method, all rules have their respective values which then will be calculated to look for the average score which then was used to determine which decision to make. Before constructing the fuzzy rules, determining the design of the game and the design of the map must be made first. After all are done, the appropriate fuzzy rules can be constructed.

The results of the survey showed that the game that uses the Tsukamoto fuzzy inference method provides an adaptive decision according to the conditions at a moment in a game. Ninth of the players who played the game agreed that game A and game B had different decisions despite using the same AI as much as 77.2%, while those who said that AI was the same as much as 22.8%. From the results of this survey it can be said that the Tsukamoto fuzzy inference method can produce adaptive decisions based on the condition of the game at a certain moments.