Reinforcement Learning in MATLAB and Simulink

Day 1 of 1

Environment and Rewards

Objective:Set up an environment and shape rewards in Simulink or MATLAB.

• Set up environment in Simulink
• Write a reward function
• Set up an agent using Simulink and MATLAB
• Connect agent and environment

Policy and Agent

Objective:Create an policy representation and construct an agent.

• Represent a policy with a neural network
• Create a reinforcement learning agent in MATLAB
• Specify simulation options to run a simulation

Neural Networks and Training

Objective:Assemble a neural network for a policy representation and train an agent.

• Assemble a neural network
• 深层网络设计师应用
• Training an agent
• Reinforcement Learning Designer app

Deployment

Objective:Generate code from a trained agent.

• Generate code
• Validation of code