February 2020
Intermediate to advanced
328 pages
8h 19m
English
The MDPtoolbox package also provides an implementation of the value iteration algorithm so that we can solve an MDP. The following code block demonstrates the same:
mdp_value_iteration(P=actions, R=rewards, discount = 0.2)
The following screenshot displays the optimal policy details:

The optimal policy that's given by the value iteration method is {2,3,1,1}. The value at the last step is close to what we got in the policy iteration method.
Read now
Unlock full access