Research on self-learning model based on genetic algorithms with application to path tracking in CGF

A self-learning model based on genetic algorithms is put forward with application to path tracking in computer generated forces (CGF). On the basis of agent, the model is constructed to improve the autonomous performance of CGF entities under path tracking environments. First, the framework of the proposed self-learning model is presented. Second, it elaborates the realization, including the principles of condition and action parts of the rule, and the fitness function design. Finally, the parameters and the generalization ability are analyzed in detail. A visible validation system is established to verify the availability and feasibility of the presented self-learning model.