Some Technologies' Research And Application In Competitive Gaming Systems

Item response theory (IRT) is a modern psychological measurement theory which built on the basis of two concepts, the examinee's performance can be predicted by a group of potential characteristics, and the relationship between the performance and potential characteristics can be explained by a continuous function. IRT breakthrough many limitations in the classical theory, and be extensively used in various areas. We have studied IRT for more than 20 years in our country, but basically confined to the psychological testing and educational testing, and there is few research on using IRT in games. However, Competitive games and psychological tests are very similar in some ways. Therefore, the thesis applies IRT to competitive games by study competitive games for primary and secondary schools.The thesis detailed analysis of the item response theory, and focus on the logistic model. Parameters estimation is one of most important issue in Item Response Theory, and the existing parameter estimation method include maximum likelihood estimation and the Bayesian estimation have some demerits. Genetic algorithms can be applied to determine its characteristics of the general non-linear system model parameter estimation, the thesis in the basic genetic algorithm analysis on the basis of combining the three-parameter logistic model, the genetic algorithm was improved, the proposed based on improved genetic algorithm 3 PLM parameter estimation algorithm.In online games, with the increase in visits, the demands on the server will be getting higher and higher, competitive gaming systems are also faced with this problem. A better way to improve server processing performance is to use clustering, and use load balancing scheduling algorithm to balance the various servers in the cluster, making the best utilization in resources. The thesis presents a dynamic load balancing algorithm for heterogeneous cluster, based on analysis of present load balancing algorithms.Finally, the thesis analyses the needs of the competitive gaming systems for primary and secondary schools, in order to improve the playability, with OGRE engine achieve in the client, and set up an effective incentive mechanism; choose the three-parameter logistic model as a system test model, and apply the improved genetic algorithms to estimate the ability of players, then, select items suitable for his ability , and enhance learning efficiency; construction of a game server cluster, and uses dynamic load balancing algorithm for heterogeneous cluster to balance servers.