Dynamic programming

[nickumia]

HW 1

Create Gridworld environment. You define a grid world with

• [x] must have one(or multiple) start state(s) and one(or multiple) final state(s).
• [x] the size of column or row >5
• [x] must have a few blocks(obstacles)
• [x] moving out of the boundary or moving to a block cell stays in the current state

Design/Configure the following components:

• [x] Deterministic Transition
• [x] Probabilistic Transition
• [x] Policy Evaluation (prediction)
• [x] Policy Improvement (greedy)
• [x] Policy Iteration
• [x] Value Iteration

Complete the following outputs:

• Run in deterministic transition

  • [x] Given a uniform policy, show the results of policy_eval
  • [x] Run policy iteration and show the results
  • [x] Run value iteration and show the results. After that extract policy from v values

• Run in Probabilistic transition

  • [x] Given a uniform policy, show the results of policy_eval
  • [x] Run policy iteration and show the results
  • [x] Run value iteration and show the results. After that extract policy from v values

Not Required, but helpful?

• [x] Some form of GUI?
• [x] Some representation of the Policy?
• [ ] Some representation of the State Values?
• [ ] Some representation of the Action Values?
• [x] Some representation of the sample simulation?

References:

• https://www.analyticsvidhya.com/blog/2018/09/reinforcement-learning-model-based-planning-dynamic-programming/

[nickumia]

Follow-on work in

• https://github.com/nickumia/cap6629/issues/2