Algorithms Searching for Solutions in Fighting Games

In this Paper, we introduce Algorithms searching for solutions in fighting games, a common layer of problems in the field of Artificial Intelligence (AI). There are many types of fighting games. The complexity of algorithms in terms of memorial space and computational time is often very large. Each particular game has its own rules of movements, starting states and ending states. In order to describe this problem, 5 components including starting state, intermediate state, ending state, transition, and search tree are used. Finding a new move advantageously is a search tree basing on the evaluating function. From the general algorithm, in order to reduce the complexity, the paper also introduces the α-β cut method. According to this method, some branches in the search tree can be ignored. When applying the limited search depth level and the α-β cut method, the algorithm produces a faster searching result and allows the player to have multiple choices about the difficulty level of the games.