Author ORCID Identifier

Fabio Vitor

Document Type

Article

Publication Date

9-23-2017

Publication Title

Mathematical Methods of Operations Research

Volume

87

First Page

109

Last Page

137

Abstract

The simplex method, created by George Dantzig, optimally solves a linear program by pivoting. Dantzig’s pivots move from a basic feasible solution to a different basic feasible solution by exchanging exactly one basic variable with a nonbasic variable. This paper introduces the double pivot simplex method, which can transition between basic feasible solutions using two variables instead of one. Double pivots are performed by identifying the optimal basis in a two variable linear program using a new method called the slope algorithm. The slope algorithm is fast and allows an iteration of DPSM to have the same theoretical running time as an iteration of the simplex method. Computational experiments demonstrate that DPSM decreases the average number of pivots by approximately 41% on a small set of benchmark instances.

Comments

This is a post-peer-review, pre-copyedit version of an article published in Mathematical Methods of Operations Research. The final authenticated version is available online at: https://doi.org/10.1007/s00186-017-0610-4.

Included in

Mathematics Commons

Share

COinS