1: \begin{abstract}

2: In a recent paper, a distributed algorithm was proposed for solving linear algebraic

3: equations of the form $Ax = b$ assuming that the equation has at least one solution. The

4: equation is presumed to be solved by $m$ agents assuming that each agent knows a subset of the

5: rows of the matrix $\matt{A & b}$, the current estimates of the equation's solution generated by each

6: of its neighbors, and nothing more. Neighbor relationships are represented by a time-dependent

7: directed graph $\bbb{N}(t)$ whose vertices correspond to agents and whose arcs characterize neighbor

8: relationships. Sufficient conditions on $\bbb{N}(t)$ were derived under which the algorithm can cause all

9: agents' estimates to converge exponentially fast to the same solution to $Ax = b$. These

10: conditions were also shown to be necessary for exponential convergence, provided the data

11: about $\matt{A & b}$ available to the agents is ``non-redundant''. The aim of this paper is to relax

12: this ``non-redundant'' assumption. This is accomplished by establishing exponential convergence

13: under conditions which are the weakest possible for the problem at hand;

14: {\color{black} the conditions are based on a new notion of graph connectivity.

15: An improved bound on the convergence rate is also derived.}

16: \end{abstract}

17: