[Ml-stat-talks] FW: IDeAS Seminar - Wed. 3/7/12

Amit Singer amits at math.Princeton.EDU
Tue Mar 6 09:49:33 EST 2012



From: Valerie Marino [mailto:vmarino at math.princeton.edu] 
Sent: Tuesday, March 06, 2012 8:51 AM
To: PACM Grad Students
Cc: PACM etc.; pacmcertif at math.princeton.edu; Peter Constantin; Phill
Holmes; E Weinan; Amit Singer; Philippe Rigollet
Subject: IDeAS Seminar - Wed. 3/7/12


Date:  Wednesday,  March 7, 2012

Time:  3:00 pm

Place:  214 Fine Hall

Seminar:  IDeAS Seminar

Speaker: George F. Young, Mechanical & Aerospace Engineering, Princeton

Title: Optimising Robustness of Consensus to Noise
Abstract: One of the key benefits of consensus in a multi-agent system is
that it can improve decision making in the presence of uncertainty. However,
the act of communication itself will usually introduce additional noise to
the system. For this reason it is important to understand how well a group
can reach and maintain consensus in the presence of communication noise and
in particular whether it is possible to minimise the disturbances created by
this noise. Here we use a measure of group dispersion due to noise, which
can be interpreted as the H2 norm of the consensus system, to relate the
communication graph between the agents to the group robustness. We explore
some fundamental bounds on performance and investigate how to improve
robustness when the graph is a tree. Our metric is also used to evaluate
different sensing strategies in starling flocks to determine which ones
achieve robust consensus in the most efficient manner.
Valerie Marino
Program Secretary
Program in Applied & Computational Mathematics
Tel: 609-258-3703
Fax: 609-258-1735
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cs.princeton.edu/pipermail/ml-stat-talks/attachments/20120306/2740437a/attachment.html>

More information about the Ml-stat-talks mailing list