Novel Model-free Optimal Active Vibration Control Strategy Based on Deep Reinforcement Learning

Algorithm 1

DRL controller training.

	Initialization step: initial DDPG agent, initial state vector
	repeat
	Observe state and get action
	Execute in the environment
	Observe the next state
	Get the reward
	If is terminal, then reset the initial states and start a new episode
	If it is time to update, then
	DDPG
	end if
	until convergence