An event-triggered distributed primal-dual algorithm for Network Utility maximization

Many problems associated with networked systems can be formulated as network utility maximization (NUM) problems. This paper presents a distributed primal-dual algorithm for the NUM problem that uses event-triggering. Under event triggering, each agent broadcasts to its neighbors when a local ¿error¿ signal exceeds a state dependent threshold. The paper establishes such state-dependent event-triggering thresholds under which the proposed algorithm converges. The paper gives an upper bound on the largest number of successive data dropouts the network can tolerate while ensuring the algorithm's convergence. Simulation results show that the proposed algorithm reduce the number of message exchanges by up to two orders of magnitude, and enjoys much better scalability with respect to the above two measures of network size than commonly used dual decomposition algorithms.