Deep Deterministic Policy Gradient-based intelligent control scheme design for DC-DC circuit

The DC-DC circuit studied in this paper is widely used in power management to obtain the desired output voltage. However, there are factors such as non-linear characteristics and uncertain interference in the DC-DC circuit, which will cause the load voltage of the DC-DC circuit to be unstable. Aiming at the above-mentioned problems, in order to achieve fast and adaptive voltage regulation, this paper proposes a control algorithm based on the Deep Deterministic Policy Gradient (DDPG), and then use the learning ability of the DDPG controller to improve the voltage results of the converter system, so that it can use its adaptive ability to improve the voltage regulation of the DC-DC circuit, and achieve precise control of the voltage output and improve the response speed. At the same time, because multiple system target factors are considered in the evaluation function, there is no need to adjust the controller parameters, which is more intelligent. To confirm effectiveness of proposed control scheme, simulation results are tested to show that, the intelligent control algorithm can improve the charging performance of the system.