Algorithm 1From: A projected primal-dual gradient optimal control method for deep reinforcement learningAlgorithm based on NOCBack to article page