PyTorch implementation of continuous action actor-critic algorithm. The algorithm uses DeepMind's Deep Deterministic Policy Gradient DDPG method for updating the actor and critic networks along with Ornstein–Uhlenbeck process for exploring in continuous action space while using a Deterministic policy.
DDPG is a policy gradient alogrithm, that uses stochastic behaviour policy for exploration (Ornstein-Uhlenbeck in this case) and outputs a deterministic target policy, which is easier to learn.
Policy Estimation (Actor)
Actor Network consists of a 3-layer neural network taking into input the state (s) and outputs the action (a) which should be taken denoted by Pi(s).
Policy Evaluation (Critic)
Critic Network consists of a 3-layer neural network taking into input the state (s) and correspoding action (a) and outputs the state-action value function denoted by Q(s,a).
The policy is optimized by minimizing the loss :- sum ( -Q(s,a) ).
The critic is optimized by minimzing the loss :- L2( r + gamma*Q(s1,Pi(s)) - Q(s,a) ).
The above updates however don't tend to converge according to DeepMind's paper and they hence use soft policy updates by maintaing a target actor and critic whose weights are updated after above optimizations as follows :-
target_actor = beta*actor + (1-beta)*target_actor target_critic = beta*critic + (1-beta)*target_critic
where beta = 0.001
Performance of DDPG on OpenAI Envs
Below is the performance of the model after 70 episodes. Full Video
Below is the performance of the model after 900 episodes. Full Video