No confident code implementation found for this paper.
This paper establishes that an MDP with a unique optimal policy and ergodic associated transition matrix ensures the convergence of various versions of the Value Iteration algorithm at a geometric rate that exceeds the discount factor γ for both discounted and average-reward criteria.
No confident code match yet
We couldn't find an author-owned or strongly-evidenced community implementation for this paper. Any repos shown below are weak matches — verify before relying on them.
No code implementations found yet.
Know of an implementation? Let us know in the comments below!
Geometry-Inspired Unified Framework for Discounted and Average Reward MDPs
A Computationally Efficient Algorithm for Infinite-Horizon Average-Reward Linear MDPs
On the Global Convergence of Policy Gradient in Average Reward Markov Decision Processes
MDP Geometry, Normalization and Reward Balancing Solvers
Optimal Convergence Rate for Exact Policy Mirror Descent in Discounted Markov Decision Processes
Reinforcement Learning for Infinite-Horizon Average-Reward Linear MDPs via Approximation by Discounted-Reward MDPs
@article{mustafin2024value,
title = {On Value Iteration Convergence in Connected MDPs},
author = {Arsenii Mustafin and Alex Olshevsky and Ioannis Ch. Paschalidis},
year = {2024},
eprint = {2406.09592},
archivePrefix = {arXiv},
url = {http://arxiv.org/abs/2406.09592v1}
}Shows whether official code is linked for this paper.
[](https://codeofpaper.com/paper/2406.09592)