Dadush, Daniel, Végh, László A. ORCID: 0000-0003-1152-200X and Zambelli, Giacomo
(2020)
*Rescaling algorithms for linear conic feasibility.*
Mathematics of Operations Research, 45 (2).
732 - 754.
ISSN 0364-765X

Text (Coordinate_descent)
- Accepted Version
Download (626kB) |

## Abstract

We propose simple polynomial-time algorithms for two linear conic feasibility problems. For a matrix A ∈ R m× n, the kernel problem requires a positive vector in the kernel of A, and the image problem requires a positive vector in the image of A T. Both algorithms iterate between simple ﬁrst-order steps and rescaling steps. These rescalings improve natural geometric potentials. If Gofﬁn's condition measure ρ A is negative, then the kernel problem is feasible, and the worst-case complexity of the kernel algorithm is O((m 3n + mn 2)log|ρ A| −1); if ρ A > 0, then the image problem is feasible, and the image algorithm runs in time O(m 2n 2 log ρ A −1). We also extend the image algorithm to the oracle setting. We address the degenerate case ρA = 0 by extending our algorithms to ﬁnd maximum support nonnegative vectors in the kernel of A and in the image of A T. In this case, the running time bounds are expressed in the bit-size model of computation: for an input matrix A with integer entries and total encoding length L, the maximum support kernel algorithm runs in time O((m 3n + mn 2)L), whereas the maximum support image algorithm runs in time O(m 2n 2L). The standard linear programming feasibility problem can be easily reduced to either maximum support problems, yielding polynomial-time algorithms for linear programming.

Item Type: | Article |
---|---|

Official URL: | https://pubsonline.informs.org/journal/moor |

Additional Information: | © 2020 INFORMS |

Divisions: | Mathematics |

Subjects: | Q Science > QA Mathematics |

Date Deposited: | 13 May 2019 13:24 |

Last Modified: | 20 Oct 2021 00:56 |

URI: | http://eprints.lse.ac.uk/id/eprint/100778 |

### Actions (login required)

View Item |