Self-adaptive interference resource allocation method based on perception enhancement network

The invention discloses a self-adaptive interference resource allocation method based on a perception enhancement network, and belongs to the technical field of deep reinforcement learning. The perception enhancement network designed by the invention comprises a state identification network and an action value network, the state identification network is used for identifying the state value of a communication party and enhancing the perception capability of the anti-interference state of the communication party, and then the state value is fed back to the action value network to select an optimal interference action; therefore, the model is helped to match the optimal interference resource allocation strategy more accurately. Moreover, a dual-objective optimization reward function is constructed, and the optimal balance between interference success and resource saving is achieved by adjusting the relative importance weight of the two normalized reward functions. In addition, a preferential experience playback