Research And Application On Key Techniques Of Game AI

There are various of game engines and technologies for computer game developers. As the basic technology of computer games and video games, computer graphics and video technology are updated annually. Whether a game is good or bad depends largely on graphics technology, but with the extremely enhanced of graphics technologies, people are no longer satisfied with a beautiful gorgeous audio-visual experience. The game turned to the pursuit of deeper meaning, so the gameplay, which is largely depends on the artificial intelligence technologies, is becoming an important part of the game design.This paper introduced one of the most popular game AI techniques, which is behavior tree, including the shared-node behavior tree, the relation between FSM and behavior tree, and the application of machine learning to the behavior tree. The paper provided a enhanced behavior tree which uses reinforcement learning. This technology can assist game AI developers to optimize the behavior tree and automate tree design.This paper also designed and implemented a common artificial intelligence engine based on behavior tree and Unity3D. This engine combines the current mainstream of game artificial intelligence technology, suitable for most types of game AI development. The overall architecture of this engine is divided into three levels:basic framework classes, the intelligent agent’s model and the message system. The engine provides basic artificial intelligence algorithms for computer games. Developers can easily use the relevant algorithms, or provide their own algorithms by implementing certain interfaces.