Solving Linear Programs in the Current Matrix Multiplication Time
Data Structures and Algorithms
2020-10-21 v3
Abstract
This paper shows how to solve linear programs of the form with variables in time where is the exponent of matrix multiplication, is the dual exponent of matrix multiplication, and is the relative accuracy. For the current value of and , our algorithm takes time. When , our algorithm takes time. Our algorithm utilizes several new concepts that we believe may be of independent interest: We define a stochastic central path method. We show how to maintain a projection matrix in sub-quadratic time under multiplicative changes in the diagonal matrix .
Cite
@article{arxiv.1810.07896,
title = {Solving Linear Programs in the Current Matrix Multiplication Time},
author = {Michael B. Cohen and Yin Tat Lee and Zhao Song},
journal= {arXiv preprint arXiv:1810.07896},
year = {2020}
}
Comments
STOC 2019, JACM 2020