Research on Behavior Mathematical Modeling of CGF Based on Reinforcement Learning under Complex Electromagnetic Environment

This paper focuses on the modeling of agent behavior in complex electromagnetic environment. By using reinforcement learning technology, The traditional Markov decision process and dynamic programming algorithm are improved. The electromagnetic field intensity function is integrated with Markov decision process and dynamic programming algorithm, and high precision virtual battlefield environment is evaluated based on system simulation, in order to facilitate the modeling of the agent’s behavior, this paper meshes the complex electromagnetic battlefield environment and sets up the simulation task scene of agent’s maneuvering task using simulation and evaluation system. Several simulation tests were implemented to collect relative test data. The comparison was made with the traditional Markov decision algorithm from three dimensions, including time, space and working condition of electronic devices inside equipment. The results show that the improved Markov decision algorithm and dynamic programming algorithm proposed in this paper show better adaptability to adapt complex electromagnetic environment than traditional Markov algorithm, it will provide a strong support for efficient completion of combat tasks and improvement of combat effectiveness in complex electromagnetic simulation environment, and also lays a foundation for the subsequent research on the CGF behavior modeling method.