A Jamming Decision-Making Method Based on Next State Prediction SAC

In light of the growing complexity of the electromagnetic environment, traditional jamming decision-making methods have become less effective due to poor generalization and inefficient jamming performance. This paper proposes a jamming decision-making method based on next state prediction SAC (Soft Actor-Critic) to address this issue. Firstly, we design the state, action, and reward elements under electronic warfare scenarios. Secondly, a next state prediction critic network is proposed to accurately estimate the value function by modeling the environment dynamics and extracting richer representational information. Finally, a negative activation ReLU is introduced to improve the utilization ratio of network parameters. Simulation results show that the proposed method achieves a 98.4% success rate at a distance of 90 km from the radar and a 70.0% success rate at 30 km. Therefore, this method exhibits good adaptability and jamming performance in the field of jamming decision-making.