Train AC Agent to Balance Cart-Pole System
This example shows how to train an actor-critic (AC) agent to balance a cart-pole system modeled in MATLAB®.
For more information on AC agents, see Actor-Critic Agents. For an example showing how to train an AC agent using parallel computing, see Train AC Agent to Balance Cart-Pole System Using Parallel Computing.
Cart-Pole MATLAB Environment
The reinforcement learning environment for this example is a pole attached to an unactuated joint on a cart, which moves along a frictionless track. The training goal is to make the pendulum stand upright without falling over.
For this environment:
The upward balanced pendulum position is
0radians, and the downward hanging position is
The pendulum starts upright with an initial angle between –0.05 and 0.5 rad.
The force action signal from the agent to the environment is from –10 to 10 N.
The observations from the environment are the position and velocity of the cart, the pendulum angle, and the pendulum angle derivative.
The episode terminates if the pole is more than 12 degrees from vertical or if the cart moves more than 2.4 m from the original position.
A reward of +1 is provided for every time step that the pole remains upright. A penalty of –5 is applied when the pendulum falls.
For more information on this model, see Load Predefined Control System Environments.
Create Environment Interface
Create a predefined environment interface for the pendulum.
env = rlPredefinedEnv("CartPole-Discrete")
env = CartPoleDiscreteAction with properties: Gravity: 9.8000 MassCart: 1 MassPole: 0.1000 Length: 0.5000 MaxForce: 10 Ts: 0.0200 ThetaThresholdRadians: 0.2094 XThreshold: 2.4000 RewardForNotFalling: 1 PenaltyForFalling: -5 State: [4x1 double]
env.PenaltyForFalling = -10;
The interface has a discrete action space where the agent can apply one of two possible force values to the cart, –10 or 10 N.
Obtain the observation and action information from the environment interface.
obsInfo = getObservationInfo(env); actInfo = getActionInfo(env);
Fix the random generator seed for reproducibility.
Create AC Agent
An AC agent approximates the long-term reward, given observations and actions, using a critic value function representation. To create the critic, first create a deep neural network with one input (the observation) and one output (the state value). The input size of the critic network is 4 since the environment has four observations. For more information on creating a deep neural network value function representation, see Create Policies and Value Functions.
criticNetwork = [ featureInputLayer(4,'Normalization','none','Name','state') fullyConnectedLayer(32,'Name','CriticStateFC1') reluLayer('Name','CriticRelu1') fullyConnectedLayer(1, 'Name', 'CriticFC')]; criticNetwork = dlnetwork(criticNetwork);
Specify options for the critic representation using
criticOpts = rlOptimizerOptions('LearnRate',1e-2,'GradientThreshold',1);
Create the critic representation using the specified deep neural network. You must also specify the action and observation information for the critic, which you obtain from the environment interface. For more information, see
critic = rlValueFunction(criticNetwork,obsInfo);
An AC agent decides which action to take, given observations, using an actor representation. To create the actor, create a deep neural network with one input (the observation) and one output (the action). The output size of the actor network is 2 since the environment has two possible actions, –10 and 10.
Construct the actor in a similar manner to the critic. For more information, see
actorNetwork = [ featureInputLayer(4,'Normalization','none','Name','state') fullyConnectedLayer(32, 'Name','ActorStateFC1') reluLayer('Name','ActorRelu1') fullyConnectedLayer(2,'Name','ActorStateFC2') softmaxLayer('Name','actionProb')]; actorNetwork = dlnetwork(actorNetwork); actorOpts = rlOptimizerOptions('LearnRate',1e-2,'GradientThreshold',1); actor = rlDiscreteCategoricalActor(actorNetwork,obsInfo,actInfo);
To create the AC agent, first specify the AC agent options using
agentOpts = rlACAgentOptions(... 'ActorOptimizerOptions',actorOpts, ... 'CriticOptimizerOptions',criticOpts,... 'EntropyLossWeight',0.01);
Then create the agent using the specified actor representation and the default agent options. For more information, see
agent = rlACAgent(actor,critic,agentOpts);
To train the agent, first specify the training options. For this example, use the following options.
Run each training episode for at most 1000 episodes, with each episode lasting at most 500 time steps.
Display the training progress in the Episode Manager dialog box (set the
Plotsoption) and disable the command line display (set the
Stop training when the agent receives an average cumulative reward greater than 480 over 10 consecutive episodes. At this point, the agent can balance the pendulum in the upright position.
For more information, see
trainOpts = rlTrainingOptions(... 'MaxEpisodes',1000,... 'MaxStepsPerEpisode',500,... 'Verbose',false,... 'Plots','training-progress',... 'StopTrainingCriteria','AverageReward',... 'StopTrainingValue',480,... 'ScoreAveragingWindowLength',10);
You can visualize the cart-pole system during training or simulation using the
Train the agent using the
train function. Training this agent is a computationally intensive process that takes several minutes to complete. To save time while running this example, load a pretrained agent by setting
false. To train the agent yourself, set
doTraining = false; if doTraining % Train the agent. trainingStats = train(agent,env,trainOpts); else % Load the pretrained agent for the example. load('MATLABCartpoleAC.mat','agent'); end
Simulate AC Agent
simOptions = rlSimulationOptions('MaxSteps',500); experience = sim(env,agent,simOptions);
totalReward = sum(experience.Reward)
totalReward = 500