JAX Graph Optimal Transport Package
-
Updated
Mar 12, 2026 - Python
JAX Graph Optimal Transport Package
Large-scale financial collateral optimization using the Primal-Dual Hybrid Gradient (PDHG / Chambolle-Pock) algorithm implemented in JAX
Reproducible GPU/CPU benchmark evidence for Gurobi PDHG, barrier, and dual simplex on structured LP families and public LP relaxations.
This project implements a sparse Primal–Dual Hybrid Gradient (PDHG) solver for linear programs in standard form.
To associate your repository with the pdhg topic, visit your repo's landing page and select "manage topics."