Question: [5 points] Consider the specification of a Markov Decision Process according to the following figure. Code your own implementation of Value Iteration and compute the optimal policy Indicate the original utilities you used in order to start the process. Provide at least 5 inter- as well as the optimum utilities for this challenge. mediate results (in terms of optimum utilities and policies) depending on the number of iterations needed for convergence as well as the final results. Describe your implementation and your con- vergence criterion. Report computation time and number of iterations. a sj T(si, a, sj ) Si S T(S) 0.2 $1 $1 a1 0.8 C $1 a1 $2 0 $1 0.2 $2 a2 $1 $1 0.8 a2 $4 $1 0.2 0 a2 $2 0.8 $3 $2 a2 0.2 a2 : 0.8 $2 $2 0.8 a2 : 0.2 $3 $2 a3 $1 1 a3 : 0.2 $2 $3 a4 $2 1 a4 : 1 a3 SA $3 0.1 a1 SA 0.9 I : 80 S4 a1 : 0.9 S4 a1 S3 80: 80 0.2 a1 : 0.8 S4 S4 0.8 a2 : 0.8 SA a4 $1 SA a1 : 0.1 $1 a4 : 0.2 a1 : 0.2 04 : 0.8 a2 : 0.2