Multilevel Picard scheme for solving high-dimensional drift control problems with state constraints
Yuan Zhong · 2025
Read Multilevel Picard scheme for solving high-dimensional drift control problems with state constraints by Yuan Zhong (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-21607.txt.
Introduction
Yuan Zhong controls high-dimensional systems whose state must never go negative, a constraint that defines queueing networks, where waiting jobs pile up across many stations. Steering such networks well means solving a drift control problem over a finite horizon, yet exact dynamic programming collapses under the dimension of realistic systems.
His simulation-based scheme builds the value function and its gradient through multilevel Picard iterations, a recursive Monte Carlo method. The headline theorem says the method beats the curse of dimensionality. Reaching a chosen accuracy costs work that grows only polynomially with the number of dimensions and with the inverse accuracy. Numerical trials on heavy-traffic scheduling problems for parallel server systems stay feasible up to at least twenty dimensions, suggesting the approach can serve real network control.
Catalog record
| CREATOR | Yuan Zhong |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 273 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-21607