Advances in the parallelization of the simplex method

Abstract

The simplex method has been successfully used in solving linear programming problems for many years. Parallel approaches for the simplex method have been extensively studied in the literature due to the intensive computations required, especially for the solution of large linear problems (LPs). In this paper, first a detailed overview is given of the parallelization attempts concerning the standard and the revised simplex method made to date. Next, some of the most recent and significant relevant attempts are selected and presented in more detail along with experimental results. The latter include some impressive results obtained for the revised simplex method over a modern supercomputer, as well as the recent advances in GPU-based related attempts.