Batched First-Order Methods for Parallel LP Solving in MIP

Nicolas Blin, Stefano Gualandi, Christopher Maes, Andrea Lodi, Bartolomeo Stellato · 2026

Read Batched First-Order Methods for Parallel LP Solving in MIP by Nicolas Blin, Stefano Gualandi, Christopher Maes, Andrea Lodi, Bartolomeo Stellato (2026) 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-2601-21990.txt.

Description

Nicolas Blin and co-authors speed up a core routine inside mixed-integer programming. Solvers for such problems repeatedly solve related linear programs, for instance when testing branches in strong branching or tightening bounds. Solving them one by one wastes the power of modern graphics processors.

The team presents a batched first-order method that solves many linear programs in parallel on GPUs. It extends the primal-dual hybrid gradient algorithm to handle whole batches at once. Their 2026 preprint targets the inner loop of industrial solvers. Faster linear programming batches translate directly into faster logistics, scheduling, and energy planning, where mixed-integer models decide real operations every day.

Catalog record

CREATORNicolas Blin, Stefano Gualandi, Christopher Maes, Andrea Lodi, Bartolomeo Stellato
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE56 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-21990

Browse the full library · About the Tower of Babel