Algorithm 2. Train RL Agent in MDP Environment

Result: The Agent Successfully Finds The Optimal Path Which Results In Cumulative Reward

Initialization;

Create MDP Environment;

1. Create MDP Model With Identified States And Actions;

2. Specify The State Transition And Reward Matrices For The MDP;

3. Specify The Terminal States Of The MDP;

4. Create The RL MDP Environment For This Process Model;

5. Specify The Initial State Of The Agent By Specifying A Reset Function;

Create Q-Learning Agent;

1. Create A Q Table Using The Observation And Action Specifications From The MDP Environment;

2. Set The Learning Rate Of The Representation;

3. Create A Q-learning Agent;

Train Q-Learning Agent;

1. Specify The Training Options (Episode, Stop Training Criteria);

2. Train The Agent Using The ‘train’ Function;

Validate Q-Learning Results;

1. Simulate The Agent In The Training Environment Using The ‘sim’ Function.