MerEsf / Reinforcement_learning

Reinforcement Learning Projects:

1) Policy Evaluation in Cliff Walking Environment:

Model free reinforcement learning agents for prediction

Objectives:

Design an agent for policy evaluation in the Cliff Walking environment

Method:

One-step temporal difference learning, TD(0), to estimate value functions for different policies, i.e., run policy evaluation experiments

Results:

a) Optimal policy:

b) Safety policy: