Value Iteration

This assignment seeks to implement a value iteration algorithm. This algorithm will determine the optimal policy for navigating the following world:

• The state transition distribution is deterministic, so = 1.0 for all states and actions.
• Rewards for all state transitions are zero except the following:
  • (1,1) -> (2,1) = 50
  • (2,0) -> (2,1) = 100
• S3 is the terminal state
• The discount factor is 0.8

Compilation and Running

Step 1: Compile Program

cd src
make

Step 2: Run Program

cd src
make run

The generate results and answers to the questions can be found in results.txt found in the root directory