Using matrix sparsification to solve tropical linear vector equations

TL;DR

This paper introduces a novel matrix sparsification method for solving tropical linear vector equations, transforming the problem into inequalities and providing a compact solution representation with practical computational schemes.

Contribution

It presents a new approach using matrix sparsification to solve two-sided tropical linear equations, including existence conditions and solution representations.

Findings

Established existence conditions for solutions.

Derived a compact vector form of solutions.

Compared the new method with existing solutions.

Abstract

A linear vector equation in two unknown vectors is examined in the framework of tropical algebra dealing with the theory and applications of semirings and semifields with idempotent addition. We consider a two-sided equation where each side is a tropical product of a given matrix by one of the unknown vectors. We use a matrix sparsification technique to reduce the equation to a set of vector inequalities that involve row-monomial matrices obtained from the given matrices. An existence condition of solutions for the inequalities is established, and a direct representation of the solutions is derived in a compact vector form. To illustrate the proposed approach and to compare the obtained result with that of an existing solution procedure, we apply our solution technique to handle two-sided equations known in the literature. Finally, a computational scheme based on the approach to derive all solutions of the two-sided equation is discussed.