The impact of networking on distributed consensus

Date: Friday, September 18th
Time: 12:10pm (Pizza served!)
Location: Kemper Hall 1127

Dr. Sarwate
Information Theory and Applications Center, UCSD

Host: Anna Scaglione

Abstract:

Gossip algorithms are a class of decentralized solutions to the problem of achieving consensus in a network of agents. They have attracted recent research interest because they are simple and robust -- attractive qualities for wireless ad-hoc and sensor networks. Unfortunately, the standard gossip protocol converges very slowly for many popular network models. I will discuss three ways to leverage properties of the network to achieve faster convergence : routing, broadcast, and mobility.

Biography:

Dr. Anand Sarwate is a Post-Doc at the Information Theory and Applications Center at UC San Diego since September 2008. He received his PhD in 2008 at UC Berkeley, in the Wireless Foundation Center. His research interests are on network security and sensor networks.