Application Research Of Intelligent Algorithm In Mobile Phone Games Anti-Cheating

In recent years, with the rapid development of smartphone market, rich and colorful applications not only has started a new era of communication, but also has brought people a new life style integrated communication, working and entertainment. To meet people entertainment needs, mobile-phone games played an important role. With the development of mobile phone games, more and more online games on the computer has been transplanted to mobile platform, unfortunately, some online game cheating meathods reserved at meantime, such as plugins, modifier and so on.This dissertation analyzed currently common cheating methods used in mobile-phone online game and their coping countermeasures, then improved countermeasures combined with the characteristics of cheating meathods and efficiency requirements in mobile-phone games.Firstly, we designed a layout obfuscator contained code matching obfuscator specifically used in Cocos2d-x engine respectively for encoding, testing and running stages in the process of game developement. The designed obfuscator not only takes advantage of the characteristic of Cocos2d-x engine to make Javascript more desultorily, but also utilizes code matching technology to further improve the obfuscatory effect under the condition of without increasing system overhead. Secondly, we designed multistage homomorphic encryption for key data and classified the key data in the game more scientifically and effectively by testing the server and the client, and encrypted the data in different degree according to different category to reduce the unnecessary consumption of system resources as much as possible. Finally, we applied neural network to cheating detection, and modified it based on the game featureto improve the efficiency.To verify the validity of the proposed method, the dissertation implements a smartphone online game which is applicable to Android and iOS platform. It is tested to know that the proposed obfuscationis better than the traditional layout obfuscation. There is no reduce on running efficiency in the process of execution, and in some cases, even increase, such as obfuscating long variables. The multistage homomorphic encryption of key data can effectively protect critical data from the editor, and the efficiency is better than traditional methods.we can conclude that the improved neural network can effectively find the cheater.