Reproduce “Projection-based Lyapunov method for fully heterogeneous weakly-coupled MDPs” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xiangcheng Zhang, Yige Hong, Weina Wang
The authors develop an ID policy for heterogeneous weakly coupled MDPs with shared per-step budgets. It derives individualized actions from a linear program relaxation, then executes them in reassigned priority order until a budget would be exceeded. The long-run average reward gap per arm is O(1/√N), assuming known models, fixed budget structure, and uniformly bounded mixing under aperiodic-unichain ideal policies.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.