PID Accelerated Value Iteration Algorithm

                Amir-massoud Farahmand 1 2 Mohammad Ghavamzadeh 3

             Abstract                approximation of the value or action-value functions, i.e.,
    The convergence rate of Value Iteration (VI), a      Vk+1 ← T π Vk or Qk+1 ← T  Qk . For discounted MDPs,
    fundamental procedure in dynamic programming        the Bellman operator is a contraction, and standard fixed-
    and reinforcement learning, for solving MDPs can      point ite ...