Phase transition and universality for homopolymers based on stable walks.

Speaker: 

Professor Nicola Squartini

Institution: 

UCI

Time: 

Tuesday, October 14, 2008 - 11:00am

Location: 

RH 306

We consider a polymer measure based on random walks which are based on sums of iid stable random variables.
A Gibbs measure is defined which models an attraction to the origin for these walks. A phase transition occurs as the the strength of the attraction to the origin occurs.
We examine various "thermodynamic" quantities and show they are all related to each other in a simple way and exhibit universality.

A CLT for dependent variables and some homogenization problems.

Speaker: 

Professor Nikos Zygouras

Institution: 

USC

Time: 

Tuesday, April 15, 2008 - 11:00am

Location: 

MSTB 254

I will present a criterion for the validity of the central limit theorem
for a class of dependent random variables and then I will discuss some applications of
it on random, boundary homogenization problems of nonlinear PDEs such nonlinear
parabolic ones and Navier walls.

Overlap distribution in the multiple spherical SK models (joint work Michel Talagrand).

Speaker: 

Professor Dmitry Panchenko

Institution: 

Texas A&M

Time: 

Tuesday, April 29, 2008 - 11:00am

Location: 

MSTB 254

One possible approach to the study of the geometry of the Gibbs measure in the Sherrington-Kirkpatrick
type models (for example, the chaos and ultrametricity problems) is based on the analysis of the free energy
on several replicas of the system under some constraints on the distances between replicas. In general, this
approach runs into serious technical difficulties, but we were able to make some progress in the setting of the
spherical p-spin SK models where many computations become more explicit.

Negotiations in the network bargaining problem

Speaker: 

Professor Alan Hammond

Institution: 

Courant Institute

Time: 

Thursday, May 29, 2008 - 11:00am

Location: 

MSTB 254

Abstract: consider a finite graph, with an actor sitting at each node, and a
dollar on each edge. Negotiations will be conducted between pairs of
adjacent actors over splitting the dollar on the intervening edge.
At the end of negotiations, each actor may sign at most one contract with a
neighbour, agreeing on some possibly uneven split of the dollar.
How much money is each actor likely to receive? And which matchings of the
graph are likely to arise?
Kleinberg and Tardos analysed the limiting answer - a balanced solution -
that arises from assuming that actors iteratively revise current deals using
Nash bargaining, taking the best alternative deal currently available as a
backup.

Most of the talk will be expository, I'll explain the concepts of Nash bargaining and balanced solution. If there is time, I will discuss
the rate of
convergence to the balanced solution of this type of negotiation.

Two results about large random matrices.

Speaker: 

Professor Amir Dembo

Institution: 

Stanford University

Time: 

Tuesday, April 22, 2008 - 1:00pm

Location: 

MSTB 254

We show that the properly scaled spectral measures
of symmetric Hankel and Toeplitz matrices of size N by N generated by
i.i.d. random variables of zero mean and unit variance converge weakly
in N to universal, non-random, symmetric
distributions of unbounded support, whose moments are
given by the sum of volumes of solids related to Eulerian numbers.
The universal limiting spectral distribution for
large symmetric Markov matrices
generated by off-diagonal i.i.d. random variables
of zero mean and unit variance, is more explicit, having
a bounded smooth density given by the free convolution of the
semi-circle and normal densities.

Time permitting, I will also explain the formula for the
large deviations rate function
for the number of open path of length k in random graphs
on N>>1 vertices with
each edge chosen independently with probability 0

Pages

Subscribe to RSS - Combinatorics and Probability