1: \begin{abstract}

2:   We consider a network of agents whose objective is for the aggregate

3:   of their states to converge to a solution of a linear program in

4:   standard form.  Each agent has limited information about the problem

5:   data and can communicate with other agents at discrete time instants

6:   of their choosing. Our main contribution is the synthesis of a

7:   distributed dynamics and a set of state-based rules, termed

8:   triggers, that individual agents use to determine when to

9:   opportunistically broadcast their state to neighboring agents to

10:   ensure asymptotic convergence to a solution of the linear program.

11:   Our technical approach to the algorithm design and analysis

12:   overcomes a number of challenges, including establishing convergence

13:   in the absence of a common smooth Lyapunov function, ensuring that

14:   the triggers are detectable by agents using only local information,

15:   accounting for asynchronism in the state broadcasts, and ruling out

16:   various causes of arbitrarily fast state broadcasting. Various

17:   simulations illustrate our results.

18: \end{abstract}

19: