BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Delay Reduction for Adaptive Scheduling in Wireless Networks and A
 ggregated Equilibrium for Resource Sharing Games - Loc Bui\, Stanford Univ
 ersity
DTSTART:20110413T090000Z
DTEND:20110413T100000Z
UID:TALK30635@talks.cam.ac.uk
CONTACT:Microsoft Research Cambridge Talks Admins
DESCRIPTION:This talk consists of two parts:\n\nIn the first part\, I will
  talk about delay reduction for the back-pressure algorithm. The back-pres
 sure algorithm is a well-known adaptive scheduling and routing algorithm t
 hat achieves maximum throughput in wireless networks. However\, its delay 
 performance may be quite poor even when the traffic load is not close to n
 etwork capacity due to the following two reasons. First\, each node has to
  maintain a separate queue for each commodity in the network (per-flow or 
 per-node queues)\, and only one queue is served at a time. Second\, the ba
 ck-pressure routing algorithm may route some packets along very long route
 s. We present solutions to address the above issues\, and hence\, dramatic
 ally improve its delay performance. The suggested solutions also allow eac
 h node to maintain only per-neighbor queues. This is joint work with E. At
 hanasopoulou\, T. Ji\, R. Srikant\, and A. Stolyar.\n\nIn the second part\
 , I will talk about the concept of aggregated equilibrium for resource sha
 ring games. We study a priority service model where a single server alloca
 tes its capacity to agents in proportion to their payment to the system\, 
 and users act to minimize the sum of their cost for processing delay and p
 ayment. Calculation of exact Nash equilibrium proves to be difficult\, bec
 ause both the queueing and strategic interactions introduce significant co
 mplexity into characterization of agents' processing times. To overcome th
 is complexity\, we first develop a novel approximation to the processing t
 ime of an agent that is provably optimal when the system is in heavy traff
 ic\, and performs well even away from heavy traffic. Then\, using this app
 roximation\, we introduce a novel notion of equilibrium that we call aggre
 gated equilibrium (AE). In an AE\, a single user optimizes against an aggr
 egate representation of the rest of the population\; a consistency check t
 hen requires that this aggregate representation arises from individual use
 rs' decisions. We show that AE can be found in closed form\, and serves as
  a good approximation to Nash equilibrium behavior. This is joint work wit
 h R. Johari and Y. Wu.
LOCATION:Large lecture theatre\, Microsoft Research Ltd\, 7 J J Thomson Av
 enue (Off Madingley Road)\, Cambridge
END:VEVENT
END:VCALENDAR