Model-basedSafeDeepReinforcementLearningviaa ConstrainedProximalPolicyOptimizationAlgorithm

During initial iterations of training in most Reinforcement Learning (RL) algorithms, agents perform asignificant number ofrandom exploratory steps.