Uniqueness of integer solution of linear equations

Abstract

Abstract. We consider the system of m linear equations in n integer variables Ax = d and give sufficient conditions for the uniqueness of its integer solution x ∈ {−1, 1} n by reformulating the problem as a linear program. Uniqueness characterizations of ordinary linear programming solutions are utilized to obtain sufficient uniqueness conditions such as the intersection of the kernel of A and the dual cone of a diagonal matrix of ±1's is the origin in R n . This generalizes the well known condition that ker(A) = 0 for the uniqueness of a non-integer solution x of Ax = d. A zero maximum of a single linear program ensures the uniqueness of a given integer solution of a linear equation.

Probability of unique integer solution to a system of linear equations

“…This problem, which has also been studied in [12], can be considered a generalization of the classical knapsack feasibility problem [11,7,4] of finding an n-dimensional binary integer vector y ∈ {0, 1} n such that:…”

Analytics and Optimization Techniques on Feeder Identification in Smart Grids

“…In that sense in [15] and [16] it is proved that the probability that a linear relaxation system of an integer one of the form given in (3) returns a unique integer solution, increases and approaches to 1 when the relationship between the number of constraints (measurements, m = |T|) and the number of variables (customers, n = |I|), or m/n ratio, is greater than a given threshold.…”

Inferring connectivity model from meter measurements in distribution networks

“…Mangasarian et al [18,19] have showed that when m > n/2, a system of the form (6) has a unique solution ∈ [−1, 1] n with high probability (w.h.p). For our problem, this implies that as m crosses n/2, (6) will start to retrieve the true underlying solution to the lc problem w.h.p.…”