What did I do?
- Created a stochastic simulation (gym) environment for arbitrary series-parallel system's maintenance planning at fixed regular intervals
- Weibull failure model for each subsystem
- possible actions for each subsystem are -> do nothing, repair, preventive replacement and corrective replacement
- Implemented PPO and A2C stable-baselines3 RL agents for performing the maintenance planning for the created gym-env
Note : MLP archtecture optimization and Hyperparameter tuning is not done yet