2000 character limit reached
Complexity Estimates for Fourier-Motzkin Elimination (1811.01510v2)
Published 5 Nov 2018 in cs.SC
Abstract: In this paper, we propose a new method for removing all the redundant inequalities generated by Fourier-Motzkin elimination. This method is based on an improved version of Balas' work and can also be used to remove all the redundant inequalities in the input system. Moreover, our method only uses arithmetic operations on matrices and avoids resorting to linear programming techniques. Algebraic complexity estimates and experimental results show that our method outperforms alternative approaches, in particular those based on linear programming and simplex algorithm.