1: \begin{abstract}

2: This paper studies distributed stochastic approximation algorithms

3: based on broadcast gossip on communication networks represented by

4: digraphs. Weak convergence of these algorithms is proved, and an

5: associated ordinary differential equation (ODE) is formulated

6: connecting convergence points with local objective functions and

7: network properties. Using these results, a methodology is proposed for

8: network design, aimed at achieving the desired asymptotic behavior at

9: consensus. Convergence rate of the algorithm is also analyzed and

10: further improved using an attached stochastic differential equation.

11: Simulation results illustrate the theoretical concepts.

12: \end{abstract}

13: