Design And Research Of Intelligent Agent Decision System Based On ID3 Algorithm

The intelligent agent in the game system is responsible for the emotional interaction with the player directly. The decision system of the game system is the technical achievement of "artificial intelligence" , which is directly related to the success of the game design. But the current mainstream widely recognized intelligent agent decision-making system are based on prefabricated logic to build, and their own dynamic adjustment capacity is limited, it is easier.for players to find flaws to reduce decision-making intelligence, thus affecting the user experience. The use of traditional artificial intelligence technology, intelligent agent decision-making system, although able to provide independent decision-making intelligence, but because of the use of system computing capacity and the use of more restrictions and other reasons, and can not be widely used by the market. Therefore, it is very important to develop the decision-making system which can inherit the traditional advantages of the traditional decision-making system and combine the new artificial intelligence technology with real-time and autonomous dynamic adjustment according to the running situation of the game.In this dissertation, the current mainstream and new intelligent agent decision system: fuzzy logic, artificial neural network, genetic algorithm, game script, finite state machine, behavior tree were introduced and analyzed. In view of the traditional intelligent agent decision-making system can not be very good in the actual use of the environment to maintain decision-making rationality,can not effectively adapt to thediversification of players operating problems. This dissertation presents an ID3 algorithm based on machine learning ID3 algorithm and behavior tree as intelligent agent decision system. It is through the collection of players and NPC interaction generated background data by ID3 algorithm analysis and guidance behavior tree decision-making, real-time adjustment of decision-making system to maintain the rationality of decision-making. In order to verify the validity of the new decision system,a NPC agent with this scheme is designed based on the Unity3d platform. Finally, the game system is built and experimented. The experimental results verify the effectiveness of the scheme.