Reproduce “Drift Control of High-Dimensional RBM: A Computational Method Based on Neural Networks” 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
Baris Ata, J. Michael Harrison, Nian Si
The authors extend a neural-network simulation method to choose state-dependent controls for reflected Brownian models with bounded drift and discounted or ergodic costs. They simulate paths, learn value and gradient functions, then use the gradient to select actions. In a 30-dimensional parallel-server test, policy cost is within 1% of the benchmark; this separable example does not establish accuracy for coupled networks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.