Model-Based Decision-Making And Implementation Of AI Engine

As the online game in-depth the entertainment lives of people, there is a high demand for the virtual experience of online game. The popular attacking-monster, passing, upgrade mode, can not attract more players in the past. Online game industry presses for more real, more challenging opponent. I think that AI technology will development rapidly in the future, such game as Black&White and Hello have let us surprised for its AI technology. Players are looking forward to the game like this. It has been proposed to use genetic algorithm and neural networks to achieve the AI Game Engine, but these two technologies is extremely complex and require significant CPU time. So these two technologies are not applicable in the fast action game world. The technology to achieve the AI Game Engine is now more popular finite-state machine (FSM). The FSM is a Model that based on rule-action. Because it is simple, intuitive, easy to modify, so it is widely used. But AI Game Engine that based on FSM usually make the unintelligence action, such as number of NPC occupy the same route, which cause path block; Monsters cannot fight with a team. This paper put forward a new implementation method of AI engine-based on decision model. It is more advanced than the traditional AI engine based on FSM. If we use this AI engine, the MC and NPC can make the best decision for different states in the game world, so the player will wrap themselves in a real game world.The thesis is organized as follows:Section 1 introduces the basic composition and development of game engine, and the status and role of AI engine in game.etc. Section 2 tells the traditional AI engine that based on FSM, and relative game example. Section 3 discuses the AI engine that based on decision model, and builds a FSM class. Section 4 compares the advantages and disadvantages of the two methods. It is important to elaborate the advantages of the AI engine that based on decision model. Section 5 applies the two technologies in game to compare the difference between them. Section 6 will make a summary for this paper, and make a prediction to the development of AI engine.