Safe Gaussian distributed soft actor-critic (SGDSAC) algorithm

Additionally, to facilitate the training of DRL-based energy management strategies for HEVs using Python and Simulink, we have provided a DLL-based DRL environment configuration solution. We make our approach openly available here to facilitate the reviewers' evaluation of our work.

The "log" folder contains the raw training data recorded by "TensorBoard" during algorithm training. You can convert these data into Excel files using the "LoggerVisualizer" function provided in each algorithm's code. Python environment: python 3.9.0 pytorch 2.1.2 pytorch-cuda 12.1 numpy 1.26.4 cvxopt 1.2.6 cvxpy 1.5.2 tensorboard 2.17.0 pandas 2.1.4 gym 0.26.2