********************************************************************


                                                     Seminar

             Department of Systems Engineering and Engineering Management
                                  The Chinese University of Hong Kong

------------------------------------------------------------------------------------------

　

　

　

Title

:

New Origins of Heavy Tails

　

　

　

Speaker

:

Dr. Jian Tan

　

　

IBM T. J. Watson Research

　

　

　

　

　

Date

:

May 17th, 2010 (Monday)

　

　

　

Time

:

4:30 p.m. - 5:30 p.m.

　

　

　

Venue

:

Room 513

　

　

William M.W. Mong Engineering Building

　

　

(Engineering Building Complex Phase 2)

　

　

CUHK

　

　

　

------------------------------------------------------------------------------------------

Abstract:
　

Heavy tails have been repeatedly observed in a variety of socioeconomic, biological, and technological areas, since the early studies of Pareto in 1897, and Zipf in 1949.

Inspired by the generality of the central limit theorem, which explains the ubiquity of the Gaussian distribution, in this research we investigate two new universal laws that could result in heavy tails. First, we show that the widely used retransmission/restart mechanism can cause heavy-tailed delays (and possibly zero throughput) even if the system parameters are all light-tailed. In the context of information networks, this is the first body of work that attributes heavy tails entirely to retransmission-based protocol designs. Second, we propose the modulated branching processes to study the power laws that arise in the proportional growth systems. Informally, the interpretation of our main results suggests that alternating periods of growth and reduction, e.g., economic expansions and recessions, are primarily responsible for the appearance of power law distributions.


-------------------------------------------------------------------------------------------

Biography:
　

Dr. Jian Tan received his PhD degree in Electrical Engineering from Columbia University in 2009. Then he spent one year as postdoc at the Ohio State University. He has recently joined IBM T.J. Watson Research in the System Analysis and Optimization research group.

His main research interests include applied probability, queueing theory, heavy-tailed distributions, scheduling algorithms with application in information networks and service systems.


************************* ALL ARE WELCOME ************************

　

　

　

Host

:

Prof. Li Duan

Tel

:

(852) 2609-8316/8323

Email

:

dli@se.cuhk.edu.hk

　

　

　

Enquiries

:

Prof. Nan Chen or Prof. Sean X. Zhou

　

:

Department of Systems Engineering and Engineering Management

　

　

CUHK

Website

:

http://www.se.cuhk.edu.hk/~seg5810

Email

:

seg5810@se.cuhk.edu.hk

　

　

　

********************************************************************