Decision-Making and Parameter Optimization of Anti-Jamming Measures based on HPPO

In the current dynamic jamming scenarios where radar faces diverse types and varying parameters of interference, it often requires simultaneous decision-making on a limited number of anti-jamming measures and their corresponding continuous parameters. To address this mixed action space optimization problem, this paper proposes a method for anti-jamming measure decision-making and parameter optimization based on the Hybrid Proximal Policy Optimization (HPPO) algorithm. Building on the PPO algorithm, the single output layer of the actor network is modified into two independent parallel output layers, each independently calculating the importance sampling values for measures and parameters, followed by policy updates using a shared advantage function. Simulations verify the effectiveness of the proposed algorithm.