6. 1
The Q-tensor Theory of Liquid Crystals
John M. Ball
Oxford Centre for Nonlinear PDE, Mathematical Institute, University of Oxford, UK
E-mail: ball@maths.ox.ac.uk
The lecture will survey what is known about the mathematics of the de Gennes Q-
tensor theory for describing nematic liquid crystals. This theory, despite its popularity
with physicists, has been little studied by mathematicians and poses many interesting
questions. In particular the lecture will describe the relation of the theory to other theories
of liquid crystals, specifically those of Oseen-Frank and Onsager/Maier-Saupe.
Mathematical Aspects of Infrastructure Planning
Martin Grotschel
¨
Konrad-Zuse-Zentrum f¨ r Informationstechnik Berlin, Takustraße, Berlin, Germany
u
E-mail: groetschel@zib.de
The costs of establishing, maintaining, and extending a country’s infrastructure such
as highway, railroad, public transport, pipeline, power, water, or telecommunication net-
works are immense. The utilization of mathematics in these processes ranges consider-
ably from one infratsructure area to another and from country to country. In this talk I
will address some of the planning and operational issues arising in infrastructure planning
in Germany, and I will demonstrate how mathematics can be employed to solve some of
the problems coming up.
Hamilton-Jacobi Theory in Nonholonomic Dynamics
Manuel de Leon
´
Instituto de Ciencias Matematicas, Consejo Superior de Investigaciones Cient´
ificas, Madrid, Spain
E-mail: mdeleon@imaff.cfmac.csic.es
The Hamilton-Jacoby theory is a useful and classical tool to integrate the equations of
motion in Mechanics. There were many attempts in the last decades to extend this theory
7. 2
to nonholonomic mechanics. In this lecture we will discuss the geometric framework of
the Hamilton-Jacobi theory and show how it permits to extend it for the nonhonomic case.
Mathematical Problems of Very Large Networks
L´ szlo Lov´ sz
a´ a
Department of Computer Science of the Eotvos Lor´ nd University , Budapest, Hungary
¨¨ a
E-mail:lovasz@cs.elte.hu
Suppose that we have a huge graph (we don’t even know its size), and the only way
to obtain information about it is to draw a sample of the node set of bounded size. What
properties of the graph can be deduced from this sample? What should we mean by an
answer to an algorithmic question like finding the conencted components, if we cannot
even list all nodes? What does it mean if two graphs are “close” in the sense that they
cannot be distinguished by such tests? How to model such huge graphs, and how to
approximate them by smaller ones? The graph property testing model was first introduced
by Goldreich, Goldwasser and Ron (but related questions were considered before). In the
context of dense graphs, a very general result is due to Alon and Shapira, who proved that
every hereditary graph property is testable.
In this language, the Regularity Lemma of Szemeredi states that every graph can be
approximated by a weighted graph with k nodes so that the error tends uniformly to 0 as k
tends to infinity. In the other direction, the theory of graph limits yields an approximation
by a 2-variable measurable function. which allows us to use tools from analysis. This
analytic version allows for simpler formulation of many graph theoretic problems, and
leads to various characterizations of testable properties.
We survey these results, along with analogous results for graphs with bounded degree.
Mathematics and Internet Information Restirval
Zhiming Ma
Institute of Applied Mathematics, Academy of Math and Systems Science, Beijing, China
E-mail: mazm@amt.ac.cn
In this public talk I shall briefly review some of our recent joint work (in collaboration
with Microsoft Research Asia) concerning Internet Information Retrival. I shall tell you
how can we construct a Markov process describing Web users’ real browsing behaviors,
8. 3
and thus yield an algorithm for page importance computation (referred to as BrowseR-
ank). The algorithm is likely to be a competitor to the conventional ParRank used in
Google search. Our work reveals that mathematics is becoming more and more important
in information retrival.
Functors of Infinitely Near Points on an Algebraic Surface
Ragni Piene
Centre of Mathematics for Applications, Dept. of Maths., University of Oslo, Norway
E-mail: ragnip@math.uio.no
In joint work with Steven Kleiman we study sequences of infinitely near points of a
family of algebraic surfaces that are associated to given Enriques diagrams and show
that they form a functor. The functor is representable by a smooth scheme, which maps
naturally to the Hilbert scheme of the family. I will focus on applications of this theory to
the ”Gromov-Witten” problem of enumerating singular curves on a surface.
The Normal Quotient Philosophy for Edge-transitive Graphs
Cheryl E. Praeger
School of Mathematics and Statistics, University of Western Australia, Australia
E-mail: praeger@maths.uwa.edu.au
Studying normal quotients has proved an effective way to describe the structure of many
families of finite edge-transitive graphs. The normal quotient approach was initiated in
my investigation of s-arc transitive graphs, and then refined in collaboration with Giudici
and Li to develop our theory of locally s-arc transitive graphs. I will attempt to present
the essence of this philosophy with reference to a new analysis of an infinite family of
edge-transitive graphs.
From Splines to the Index Theorem
Claudio Procesi
Dipartimento di Matematica “Guido Castelnuovo”, Sapienza Universit` di Roma, Italia
a
E-mail: procesi@mat.uniroma1.it
9. 4
The celebrated index theorem of Atiyah–Singer led to a theory of transversally elliptic
operators, with respect to some compact group of symmetries. In that case an explicit
computation of the equivariant K−theory and the associated index for linear representa-
tions of tori exists only in an implicit form. Some rather surprising connection arose with
the work of Dahmen– Micchelli on partition functions and the box–spline which allows
us to make the constructions extremely explicit (joint work with Corrado De Concini and
Michele Vergne).
Extending Some Results of Tunnell and Saito
Rajat Tandon
Department of Mathematics and Statistics, University of Hyderabad, Hyderabad, India
E-mail:rtsm@uohyd.ernet.in
One of the problems representation theorists often look at is the following: Given a
topological group G, an irreducible “representation” π of G and a closed subgroup H
how does π|H decompose into irreducibles, if at all. Tunnel looked at this problem when
G =GL(2,F), where F is a nonarchimedian local field with residue characteristic not two,
√
π is an irreducible irreducible representation of G and H = K ∗ = K − {0}, K = F( d)
being a quadratic extension of F, embedded in G under the embedding
a bd
√
a+b d → .
b a
His answer was that the multiplicity of a character χ of K ∗ whose restriction to the cen-
tre of G is equal to the central character ωπ of π in π|K∗ is given by 1 + (Π ⊗ χ−1 , ψ0 ) /2
where Π is the base change lift of π to GL(2,K), ψ0 is a nontrivial additive character of K
trivial on F and (Π ⊗ χ−1 , ψ0 ) is the corresponding epsilon factor (local root number). It
can easily be shown that in the given circumstance (Π ⊗ χ−1 , ψ0 ) = ±1 so either χ does
not occur in πK∗ or it occurs with multiplicity one. Saito gave a residue characteristic free
proof of the same result.
It was apparent from the work of Tunnel that the problem posed a challenge only in
the case that π is supercuspidal ,i.e., its coefficients have compact support modulo the
centre of G. Dipendra Prasad looked for a criterion (for χ occurring in π|K∗ ) in terms of
the local root number of the character itself. Suppose π is an irreducible supercuspidal
representation of G associated to a character θ of a separable quadratic L of F. Let G+ be
10. 5
the subgroup of index 2 in G consisting of those 2 × 2 matrices whose determinant is in
NK/F (K ∗ ). If LK then π|G+ remains irreducible but if L = K then π|G+ breaks up into two
ˆ
irreducible representations π+ and π− . If χ ∈ K ∗ occurs in π|K∗ then χ must occur in either
π+ or π− . Prasad showed that if the residue characteristic is odd then χ will occur in π+
if and only if (χθ−1 , ψ0 ) = (χσ θ−1 , ψ0 ) = 1 ( σ is the nontrivial element in the Galois
group G(K/F)) and in π− if and only if (χθ−1 , ψ0 ) = (χσ θ−1 , ψ0 ) = −1.
In a joint work with N.K. Vishnu we give a residue characteristic free proof of this
result. Saito has also given a completely different proof of this. We also count the number
of χ’s of a given conductor that occur in π|K∗ .
Solved and unsolved problems of integral geometry and
monodromy theory arising from Newton’s ”Principia”
V.A.Vassiliev
Steklov Institute of Mathematics Russian Academy of Science
Mathematics College Independent University of Moscow
E-mail: vassil@vassil.mccme.rssi.ru
Two statements from the Newtons ”Principia Mathematicae...” will be discussed: Lemma
XXVIII on the non-squarability of plane ovals, and Theorem XXXI on the attraction by
spheric layer outside the sphere. Both these results can be extended to arbitrary dimen-
sions and much more general surfaces. However, unsolved problems still exist in both
arising theories. Here are two of them:
1. For any n, do there exist compact algebraically squarable hypersurfaces in Rn other
than the ellipsoids for odd n (which are squarable by an Archimedes’ theorem)?
2. Do there exist new examples of hyperbolic layers in Rn (besides the ones described
in the talk), whose attraction fields are single-valued (or at least algebraic) in some com-
ponents of their complements, except for the hyperbolicity domain?
TBA
Marcelo Viana
IMPA, Estrada Dona Castorina 110, 22460-320 Rio de Janeiro, Brazil
E-mail: viana@impa.br
11. 6
A
Distribution of Primes and Dynamics of the w Function
, , 210097
E-mail: ygchen@njnu.edu.cn
Let P be the set of all primes, and A3 = {pqr|p, q, r ∈ P}{p3 |p ∈ P}. For n = pqr ∈ A3 ,
let w(n) = P(p + q)P(q + r)P(r + p), where P(m) denotes the largest prime factor of m,
and let wi+1 (n) = w wi (n) , (i = 1, 2, . . .). It is known that for any n ∈ A3 , there exists
an i with wi (n) ∈ {20, 98, 63, 75}. We will talk about the properties of the shifted set
a + P of P and dynamics of the w function. In particular we confirm a conjecture of
Wushi Goldring on dynamics of the w function by employing the Green-Tao Theorem on
arithmetic progressions in primes.
Applications of Lie Theory to Finsler Geometry
, , 300071
E-mail: dengsq@nankai.edu.cn
In this talk, we will introduce some recent progress on the applications of Lie theory
to the study of Finsler geometry. In Section 1, we introduce the motivations to study
Finsler geometry using Lie theory. In section 2, we recall some basic definitions and
results on Finsler geometry. In section 3, we consider the group of isometries of a Finsler
space. In Section 4, we present some algebraic method to construct homogeneous Finsler
spaces. In section 5, we study symmetric Finsler spaces. Finally, we present some result
concerning the relationship between Finsler metrics and the degree of symmetry of a
smooth manifold.
12. 7
Quivers, Quasi-Quantum Groups and Finite Tensor Categories
, , 250100
E-mail: hualin@sdu.edu.cn
In this talk I report the project of quiver approaches to quasi-quantum groups. Firstly
I introduce the general quiver framework, then I give some classification results of finite
quasi-quantum groups, finally I apply these results to study finite tensor categories.
Anomalous Primes for Elliptic Curves over Q
with Complex Multiplication
, , 210093
E-mail: qingzhji@nju.edu.cn
In this paper, we prove that there exist elliptic curves defined over Q which possess
an infinite number of anomalous primes if and only if the corresponding quadratic poly-
nomials represent infinitely many primes. In particular, we show that a prime p is an
anomalous prime for the elliptic curve E : y2 + y = x3 − 21870x + 2489899/2 if and only
if p = 432n2 ± 108n + 7 for some integer n.
Gorenstein Projective Injective and Flat Complexes
, , 730070
E-mail: liuzk@nwnu.edu.cn
A complex C is called Gorenstein projective if there exists an exact resolution of com-
plexes . . . → P−1 → P0 → P1 → · · · such that each Pi is projective, C = Ker(P0 → P1 )
and the resolution remains exact when Hom(−, Q) is applied to it for any projective com-
plex Q. Gorenstein injective complexes are defined similarly. A complex G is called
Gorenstein flat if there exists an exact sequence of complexes · · · → F−1 → F0 → F1 →
· · · such that each Fi is flat, G = Ker(F0 → F1 ) and the sequence remains exact when
E − is applied to it for any injective complex E. We will talk some characterizations
13. 8
of Gorenstein projective, injective and flat complexes of left R-modules. Also Gorenstein
homological dimensions and Gorenstein flat covers of complexes are considered.
Composition Factors of Kac-modules for General Linear Lie
Superalgebras
, , 230026
E-mail: ycsu@ustc.edu.cn
To study finite-dimensional modules of Lie superalgebras, Kac introduced certain in-
decomposable modules, now known as Kac-modules. It is known that the Kac-module
is simple if and only if the corresponding highest weight is typical. When the highest
weight is atypical, the structure of the Kac-module has been a subject of intensive study.
In this talk, we present results on composition factors of Kac-modules for general linear
Lie superalgebras.
Linear Groups Over Semilattices and Automorphism Groups for
Gradation Shifting Lie Algebras
, ,361005
E-mail: tans@xmu.edu.cn
Under similarity a semilattice S in the Euclidean space Rn can be expressed as a union
of distinct cosets of 2Zn in Zn including the trivial coset. It is known that there is a one
to one correspondence from the set of all similarity classes of semilattices onto the set
of all isomorphic classes of extended affine root systems of type A1 . The automorphism
group of a semilattice S is the group of permutations preserving all cosets in S . In this
paper, we study the structure of the automorphism groups of semilattices in Rn . The auto-
morphism groups of all non-similar semilattices with dimension 2 and 3 are determined.
As an application, we use the automorphism group of a special semilattice to determine
the automorphism group of the gradation shifting toroidal algebra L(t1 , · · · , tn , 1), which
extends our earlier work.
14. 9
K2
, ,100049
E-mail: tanggp@gucas.ac.cn
F p Cp p G
K2 (F[G × C p ]) K2 (FG) p-
2
G p- 1 F
FG K2 K2 (FG) K2 (FG)
ΩFH → K2 (FH[t]/(tk ), (t)) k ≡ 1modp k>p H p-
1
p- H HDR (FH)
H p-
On Generalized Schur’Lemma and its Application
, , 100871
E-mail: lwang@math.pku.edu.cn
In this talk, the Schur’s lemma on the basis of centralized ring is generalized to en-
domorphsim ring of any induced module. This generalized lemma can unify some old
results and get some new results.
Homological Properties of Noetherian Hopf Algebras
, , 200433
E-mail: qswu@fudan.edu.cn
I will survey the progress on the homological aspects of noetherian Hopf algebras/quantum
groups in the talk. Some classification results of noetherian Hopf algebras with low di-
mensional GK dimension will also be discussed.
15. 10
B
-
, , 100190
E-mail: qunchen@whu.edu.cn
Spin M L
M
Legendrian cable links
, , 100871
E-mail: dingfan@math.pku.edu.cn
Hansjorg Geiges
Open and Closed string theories in Geometry
, , 100871
E-mail: fanhj@math.pku.edu.cn
The geometrical realizations of the 2D topological field theory are the quantization of
the classical geometrical theories. The Geometrical realizations of the N=2 supersymmet-
ric nonlinear sigma model and the Landau-Ginzburg model have brought new quantum
invariants to the fields of symplectic geometry and the classical singularity theory and
various duality phenomena. This big theory can be divided into two type theories: the
closed string theory and the open string theory. The closed string theories in geometry
correspond to the Gromov-Witten theory and the quantum singularity theory; The open
string theories in geometry correspond to various derived categories of branes. There are
many problems arising in understanding each geometrical theory and their mutual con-
nections. So far as I understand, those known geometrical theories are only some isolated
16. 11
points in a big 2D topological field theory. I will talk about the progress in this field and
propose some problems in which I am interested in this field.
Balanced metrics
, , 200433
E-mail: majxfu@fudan.edu.cn
A balanced metric is a hermitian metric on a complex n-dimensional manifold such
that its hermitian form ω satisfies d(ωn−1 ) = 0. The balanced metric plays an important
role in the study of non-K¨ hler complex geometry and superstring theories. In this talk I
a
will review the existence result of balanced metrics on complex manifolds k (S 3 × S 3 ) for
any k ≥ 2 in my joint work with Jun Li and Shing-Tung Yau.
Complete classification of locally strongly convex affine
hypersurfaces with parallel cubic form
, , 450052
E-mail: huzj@zzu.edu.cn
The fundamental theorem of equiaffine hypersurfaces in affine differential geometry
states that the affine Blaschke-Berwald metric h and the cubic form C uniquely determine
a non-degenerate hypersurface M in the real unimodular-affine space Rn+1 up to equiaffine
equivalence. Moreover, the classical Blaschke-Pick-Berwald theorem shows that C ≡ 0
on M if and only if M is an open part of a non-degenerate quadric. The classification of
locally strongly convex hypersurfaces with ˆ ≡ 0 is a fundamental and unsolved problem,
where ˆ ≡ 0 is the Levi-Civita connection of h. The known results are only for lower
dimensional cases (n ≤ 7).
In this talk, we will give a survey about the history of the study of this problem and give
an outline of our complete classifications of locally strongly convex hypersurfaces with
ˆ ≡ 0 for all dimensions (recent joint works with H. Li and L. Vrancken), which gives a
complete answer of the above unsolved problem.
17. 12
Amalgamations of Heegaard Splittings
, ,116024
E-mail: ffcclei@yahoo.com.cn
Let Mi be a compact 3-manifold with boundary, Fi ⊂ ∂Mi a boundary component,
i = 1, 2, and h : F1 → F2 a homeomorphism. The 3-manifold M = M1 ∪h M2 is
called an amalgamation of M1 and M2 along F = F1 = F2 , and is denoted by M =
M1 ∪F M2 . A Heegaard splitting for M1 and M2 can naturally induce a Heegaard splitting
for the amalgamated 3-manifold M, which is called the amalgamated Heegaard splitting
of the two Heegaard splittings. The study of the questions if and when the genus of
an amalgamated Heegaard splitting is degenerate has addressed much attention in recent
years. In the present talk, we will survey the results on the aspect, including some quite
new progress.
Variational problems in Geometry of Submanifolds
, , 100084
E-mail: hli@math.tsinghua.edu.cn
In this talk, we give a survey of some variational problems in geometry of subman-
ifolds, which includes recent results about r-minimal submanifolds, Wulff shape and
stability of hypersurfaces with constant (r + 1)-th anisotropic mean curvature and the
anisotropic version of the Alexandrov Theorem.
Classification of Equivariant Manifolds
, , 200433
E-mail: zlu@fudan.edu.cn
In the recent years, a new research field called “ Toric Topology” is emerging. Toric
Topology is not only related to Toric Geometry but also related to many other fields, such
as Symplectic Geometry, Combinatorics, Algebra and so on. This new field is getting
active, and can gather many different ideas and theories together. Based upon this, we
18. 13
shall consider the classification of equivariant manifolds; especially for the classification
up to (equivariant) homeomorphism and equivariant cobordism. We shall give a necessary
and sufficient condition for equivariant homeomorphism, and calculate the equivariant
cobordism groups etc..
, , 100190
E-mail: xsun@math.ac.cn
The Equivariant Noncommutative Atiyah-Patodi-Singer
Index Theorem
, , 130024
E-mail: wangy581@nenu.edu.cn
In this paper, we prove an equivariant noncommutative Atiyah-Patodi-Singer index
theorem.
, , 200092
E-mail: yhyang@tongji.edu.cn
, , 215006
E-mail: zhangyiing@gmail.com
19. 14
SL(2,C)
B.H.Bowditch
S.P.Tan Y.L.Wong Bowditch
McShane Bowditch
, ,510275
E-mail: zhenghao@mail.sysu.edu.cn
Abel
Abel
Local Gromov-Witten Invariants in Arbitrary Genera
, ,100084
E-mail: jzhou@math.tsinghua.edu.cn
We will report some results on computations of Gromov-Witten invariants of some
local Calabi-Yau 3-folds in arbitrary genera.
Geometric Measures and Geometric Inequalities
, ,400715
E-mail: zhoujz@swu.edu.cn
The isoperimetric inequality is one of the well-known geometric inequalities. One may
list many of known geometric inequalities involving global differential geometry, inte-
gral geometry, convex geometry, and functional analysis. For example, there are Fenchel
inequality, Alexandrov Fenchel inequality, Bonnesen inequality, Ros inequality and the
20. 15
Sobolev inequality in analysis. Integral geometry, originated from Buffon s needle prob-
lem, closely related to probability. Buffon, Crafton, Poincare, Santalo, Blashke and Chern
made great contribution to integral geometry. Minkowski investigates convex bodies.
Hadwiger s work on geometric measures is the milestone for convex theory. Via geo-
metric measures, integral and convex geometric analysis, we will present the recent results
and the future possible research on this beautiful and fascinating branch of mathematics.
Cannonical Metrics on Toric Manifolds
, , 100871
E-mail: xhzhu@math.pku.edu.cn
In this talk, I will discuss some geometry on toric manifolds. We will talk about exis-
tence problems of Kahler-Einstein metrics as well as Calabi’s extremal metrics on such
manifolds. We can also extend our method to toric orbifolds.
21. 16
C
Hamilton
, ,010021
E-mail: alatanca@imu.edu.cn
Fourier
Hilbert-Schmidt Sturm-Liouville
20 60
Hamilton 90
Hamilton
Fourier Sturm-Liouville
Hamilton
Hamilton
Hamilton Hamilton
Hamilton
Hamilton
, ,215008
E-mail: ylcao@suda.edu.cn
Henon-like Henon-like
22. 17
The Halpern Open Problem and Viscosity Approximation
, ,300160
E-mail: chenrd@tjpu.edu.cn
In this talk, I shall introduce Halpern open problem and viscosity iterative and some
results on viscosity approximation on fixed points, solutions of a generalized equilibrium
problem recently by our group.
, ,100871
gansb@math.pku.edu.cn
Poincare
Heat Kernels on Metric Spaces
, , 100084
E-mail: hujiaxin@mail.tsinghua.edu.cn
This talk gives some recent developments on the study of heat kernels on metric mea-
sure spaces, including some interesting fractals.
, , 050016
E-mail: cljiang@mail.hebtu.edu.cn
1978 M.Cowen R.Douglas
Cowen-Douglas
23. 18
A Generalization of the Lax-Oleinik Formula in Scalar
Conservation Law
, , 100190
E-mail: thli@math.ac.cn
For the scalar conservation law ut + f (u) x = 0, when f is uniformly convex; i.e. f (u) ≥
α > 0, the Lax-Oleinik formula provides the weak solution. We generalize this formula
to the case when f is convex and f (u) is strictly increasing. For example, f (u) = exp(u)
or u4 .(Joint with Banghe Li).
, ,730000
E-mail: wtli@lzu.edu.cn
G-
Profile
Nicholson
Lyapunov
24. 19
Lotka-Volterra
A variational problem associated with the minimal speed of
travelling waves for spatially periodic reaction-diffusion equations
, , 230026
E-mail: xliang@ustc.edu.cn
We consider the equation ut = u xx + b(x)u(1 − u), x ∈ R, where b(x) is a nonnegative
measure on R that is periodic in x. In the case where b(x) is a smooth periodic function,
it is known that there exists a travelling wave – more precisely a “pulsating travelling
wave” – with average speed c if and only if c ≥ c∗ (b), where c∗ (b) is a certain positive
number depending on b. This constant c∗ (b) is called the “minimal speed”. In this paper,
we first extend this theory by showing the existence of the minimal speed c∗ (b) for any
nonnegative measure b with period L. Next we study the question of maximizing c∗ (b)
under the constraint [0,L) b(x)dx = αL, where α is an arbitrarily given positive constant.
This question is closely related to the problem studied by mathematical ecologists in late
1980’s but its answer has not been known. We answer this question by proving that the
maximum is attained by periodically arrayed Dirac’s delta functions αL δ(x + kL).
k∈Z
Heisenberg
, , 100871
E-mail: hpliu@pku.edu.cn
Heisenberg 2 Heisenberg
Heisenberg
Heisenberg Siegel
Heisenberg
Heisenberg Heisenberg Laplace
¯
∂-Neuman H-
Heisenberg Laplace
25. 20
H- H-
L1 -
KK
1 Iwasawa H- H-
1 H- AN
50 Lichnerowicz Riemanm
Heisenberg (1)H-
Wiener Feynman-Kac Schrodinger
¨
(2)H- Schrodinger
¨ Hardy
Hardy Littlewood-
Paley Hardy Schrodinger
¨
Riesz Hardy (3) Heisenberg Laplace
Heisenberg Cauchy-Szego
¨
, , 610064
E-mail: matian56@sina.com
Global Structure of Positive Solutions for Nonlocal Boundary
Value Problems Involving Integral Conditions
, ,730070
E-mail: mary@nwnu.edu.cn
We consider the nonlinear eigenvalue problems u + λh(t) f (u) = 0, 0 < t < 1 with
1 1
u(0) = 0, u(1) = u(s)dA(s), where u(s)dA(s) is a Stieltjes integral with A is non-
0 0
decreasing and A(t) is not a constant on (0, 1); h ∈ C((0, 1), [0, ∞)) and h(t) 0 on
26. 21
any subinterval of (0, 1); f ∈ C([0, 1), [0, ∞)) and f (s) > 0 for s > 0, and f0 = f∞ =
0, f0 = lim s→0+ f (s)/s, f∞ = lim s→∞ f (s)/s. We investigate the global structure of positive
solutions by using global bifurcation techniques.
, ,230026
E-mail: wxshen@ustc.edu.cn
Julia
Rivera-Letelier
Asymptotic Limits of Compressible Euler-Maxwell System
in Plasma Physics
, ,100124
E-mail: wangshu@bjut.edu.cn
In this talk we will discuss asymptotic limit problems of compressible Euler-Maxwell
system in plasma physics. Some recent results about the convergence of compressible
Euler-Maxwell system to the incompressible Euler or e-MHD equations are given, and
some new methods or ideas are reviewed.
Persistent Homoclinic Orbits for a Spatial-dependently
Perturbed Nonlinear Schr¨ dinger Equation
o
, ,230026
E-mail: wangyi@ustc.edu.cn
The existence of homoclinic orbits for nearly integrable Hamiltonian PDEs is closely
related to Chaos. In this talk, we consider a certain diffusive perturbations of the inte-
grable nonlinear Schr?dinger equation under periodic boundary conditions. The spatial-
dependence of the damped-driven term and unboundedness of the diffusion destroy the
27. 22
invariance of the plane of constants and some geometric structures. We overcome these
difficulties and prove the existence of homoclinic orbits for the perturbed NLS. This is a
joint work with Shui-Nee Chow and Chongchun Zeng.
Diffusion Equations with Degeneracy on the Boundary
, , 130012
E-mail: wangcp@jlu.edu.cn
This talk contains two parts. In the first part, we formulate the boundary problems
of quasilinear diffusion equations with boundary degeneracy and gradient nonlinearity
and establish their well-posedness. Different from the classical theory by G. Fichera and
O. A. Oleinik for the linear equations, it is shown that on the degenerate part of the
boundary, the equations may exhibit not only hyperbolic characteristics but also parabolic
characteristics. In the second part, we study the approximate controllability of a class of
semilinear systems with boundary degeneracy. The equations may be weakly degenerate
and strongly degenerate on a portion of the boundary. We prove that the control systems
are approximate controllable and the controls can be taken to be of quasi bang-bang form.
, ,400715
E-mail: wendi@swu.edu.cn
New Besov-type Spaces and Triebel-Lizorkin-type Spaces
, ,100875
E-mail: dcyang@bnu.edu.cn
28. 23
˙ p,q
Let s, τ ∈ R and q ∈ (0, ∞]. We introduce Besov-type spaces Bs,τ (Rn ) for p ∈ (0, ∞]
˙ s,τ
and Triebel-Lizorkin-type spaces F p,q (Rn ) for p ∈ (0, ∞), which unify and generalize the
Besov spaces, Triebel-Lizorkin spaces and Q spaces. We then establish the ϕ-transform
characterization of these new spaces in the sense of Frazier and Jawerth. Using the ϕ-
˙ p,q ˙ s,τ
transform characterization of Bs,τ (Rn ) and F p,q (Rn ), we obtain their embedding and lifting
properties; moreover, for appropriate τ, we also establish the smooth atomic and molec-
˙ p,q ˙ s,τ
ular decomposition characterizations of Bs,τ (Rn ) and F p,q (Rn ). For s ∈ R, p ∈ (1, ∞) and
1
q ∈ [1, ∞) and τ ∈ [0, (max{p,q}) ], via the Hausdorff capacity, we introduce certain Besov-
˙ s,τ ˙ s,τ
Hausdor spaces BH p,q (Rn ) and Triebel- Lizorkin-Hausdorff F H p,q (Rn ) and prove that the
˙ s,τ ˙ s,τ ˙ ˙
dual spaces of BH p,q (Rn ) and F H p,q (Rn ) are just, respectively, B−s,τ (Rn ) F −s,τ (Rn ) ,where
p ,q p ,q
t denotes the conjugate index of t ∈ (1, ∞). Applications to trace theorems and bound-
edness of the pseudo-differential operators with homogeneous symbols are obtained. The
inhomogeneous versions of these results are also given.
Lorenz
, ,510640
E-mail: qgyang@scut.edu.cn
Lorenz Lorenz
Hopf Silinkov
2 Marotto
, , 430074
E-mail: yangxs@mail.hust.edu.cn
29. 24
, ,210093
E-mail: huicheng@nju.edu.cn
Newton
, ,610064
E-mail: zhangshiqing@msn.com
Newton
Boltzmann BGK
, ,430074
E-mail: xwzhang@mail.hust.edu.cn
The BGK model of the Boltzmann equation plays an important role in the kinetic theory
of rarefied gases. Some existence and uniqueness results of global solutions to its Cauchy
problem and boundary value problem were established both for large and small initial data
under various circumstances. In this talk, we first review its L1 theory established by B.
Perthame and its L∞ existence and uniqueness results due to B. Perthame, M. Pulvirenti
and S. Mischler, and some related topics as well. Then, we establish certain weighted L p
estimates of the hydrodynamical quantities and local Maxwelians for rarefied gases. By
those estimates and the L∞ results, we construct approximate solutions and prove their
uniform L p bounds. Finally, we prove the existence theorem of the L p solutions to the
Cauchy problem and establish some propagation properties of the L p moments for this
kind of solutions.
30. 25
Diffusive Expansion for Solutions of the Boltzmann Equation
in the Whole Space
, , 430072
E-mail: hhjjzhao@whu.edu.cn
This talk is concerned with the diffusive expansion for solutions of the Boltzmann
equation in the whole space. It is based on a recent work joint with Dr. Shuangqian
Liu.
p-Laplace
, , 361005
E-mail: jnzhao@xmu.edu.cn
1 p-Laplace
2 3×3 Navier-Stokes
, ,730000
E-mail: ckzhong@lzu.edu.cn
Sobolev
New Progress on Critical Point Theory with Applications
, ,100084
E-mail: wzou@math.tsinghua.edu.cn
31. 26
The following topics will be involved: 1. Multi-Bump solutions and Critical Groups;
2. Bahri-Lions perturbation problem; 3. Brezis-Nirenberg problem; 4. Perturbed Brezis-
Nirenberg problem; 5. A pure critical exponent problem.
32. 27
D
, ,100190
E-mail: clq@lsec.cc.ac.cn
1.
2. / 3.
4. Maxwell
Convergence Analysis of the Jacobi-collocation Spectralmethods
for Volterra Integral Equations with a Weakly Singular Kernel
, , 510631
E-mail: yanpingchen@scnu.edu.cn
In this talk, a Jacobi-collocation spectralmethod is developed for Volterra integral equa-
tions of the second kind with a weakly singular kernel. We use some function transfor-
mation and variable transformations to change the equation into a new Volterra integral
equation defined on the standard interval [−1, 1], so that the solution of the new equation
possesses better regularity and the Jacobi orthogonal polynomial theory can be applied
conveniently. In order to obtain high order accuracy for the approximation, the integral
term in the resulting equation is approximated by using Jacobi spectral quadrature rules.
The convergence analysis of this novel method is based on the Lebesgue constants corre-
sponding to the Lagrange interpolation polynomials, polynomials approximation theory
for orthogonal polynomials and the operator theory. The spectral rate of convergence for
the proposed method is established in the L∞ -norm and weighted L2 -norm. Numerical
results are presented to demonstrate the effectiveness of the proposed method.
33. 28
, ,510275
E-mail: lnsczy@mail.sysu.edu.cn
Fast multilevel augmentation methods for solving operator equations are developed.
Each of these methods requires availability of a multilevel decomposition of the solution
space and a projection from the solution space onto a finite dimensional subspace. A
subspace at a level is obtained from the subspace at the previous level by adding a dif-
ference subspace. Accordingly, the projection equation at the present level is obtained
by augmenting the projection equation at the previous level. Using this idea recursively,
we solve the operator equation only at an initial lower resolution level while obtain its
solution at a higher resolution level. We prove that the proposed methods require only
linear computational complexity and have the optimal convergence order. A relationship
between the proposed method and the multigrid method is discussed. Two specific fast
methods based on the Galerkin projection and the collocation projection are developed.
These algorithms are applied to solve Fredholm integral equations of the second kind,
ill-posed integral equations of the first kind, Hammerstein nonlinear equations, and two-
point boundary value problems of differential equations. Numerical results are presented
to confirm the theoretical estimates and the efficiency of the methods.
, ,410083
E-mail: xlhan@mail.csu.edu.cn
B
Implicit/Explicit Schemes for the Navier-Stokes Equations
, ,710049
34. 29
E-mail: heyn@mail.xjtu.edu.cn
In this paper, we study the stability and convergence of the implicit/explicit schemes
for the two-dimensional Navier-Stokes equations. The finite element method and spec-
tral method are applied for the spatial approximation of the velocity and pressure. The
time discretization is based on the implicit schemes for the linear terms and the explicit
schemes for the nonlinear term. Moreover, we prove the stabilities and optimal error
estimates under the corresponding stability conditions, where the schemes are almost un-
conditionally stable and convergent for the smooth initial data u0 ∈ H 2 , i.e., the time
step size τ satisfies τ ≤ C0 ; and the schemes are almost weak unconditionally stable and
convergent for the non-smooth initial data u0 ∈ H 1 , i.e., the time step size τ satisfies
τ|logh| ≤ C0 for the mesh size 0 < h < 1; and the schemes are conditionally stable for the
non-smooth initial data u0 ∈ L2 , i.e., the time step size τ satisfies τh−2 ≤ C0 .
Helmholtz
, ,100190
E-mail: hqy@lsec.cc.ac.cn
Helmholtz Maxwell
Helmholtz
Helmholtz
Maxwell
, ,510631
E-mail: liwen@scnu.edu.cn
On the Numerical Simulation of Chemical Kinetic System
35. 30
, , 100871
E-mail: tieli@pku.edu.cn
The numerical simulation of chemical kinetic system opens a new way for computa-
tional mathematics community. Mathematically it deals with the stochastic simulation
for the Q-process in the theory of probability. SSA and tau-leaping methods are two typ-
ical methods, which correspond to the exact and Euler approximations to the stochastic
differential equations. In this lecture, I will talk about some recent mathematical results
about these algorithms done by our group, which include the mathematical analysis, the
construction of new algorithms, etc. I will try to show the interesting interplay of the
traditional numerical analysis and the stochastic analysis.
, ,300401
E-mail: mathlxw@hebut.edu.cn
-
Structure Exploited Algorithms for a Nonsymmetric
Algebraic Riccati Equation Arising in Transport Theory
, ,361005
E-mail: lzlu@xmu.edu.cn
The nonsymmetric algebraic Riccati equation arising in transport theory is a special
algebraic Riccati equation whose coefficient matrices having some special structures. By
reformulating the Riccati equation, they were found that its arbitrary solution matrix is of
36. 31
a Cauchy-like form and the solution matrix can be computed from a vector form Riccati
equation instead of the matrix form that. In this talk, we review some classical-type and
Newton-type iterative methods for solving the vector form Riccati equation and present
some work under investigation.
JASMIN
, ,100088
E-mail: zeyao mo@iapcm.ac.cn
JASMIN http://www.iapcm.ac/
jasmin
JASMIN
PIC
JASMIN
Adaptive RKDG Method Using Different Indicators
, ,210093
E-mail: jxqiu@nju.edu.cn
In this presentation, we systematically investigate adaptive Runge-Kutta discontinu-
ous Galerkin (RKDG) methods for hyperbolic conservation laws with different indicators
which were based on the troubled cell indicators studied by Qiu and Shu [SIAM J. Sci.
Comput., 27 (2005), 995-1013], with an objective of obtaining efficient and reliable indi-
cators to obtain better performance for adaptive computation to save computational cost.
Both h-version and r-version adaptive methods are considered in the paper. The idea is
to first identify “troubled cells” by different troubled-cell indicators, namely those cells
37. 32
where limiting might be needed and discontinuities might appear, then adopt an adaptive
approach in these cells. A detailed numerical study in one dimensional case is performed,
addressing the issues of CPU cost, accuracy, non-oscillatory property, and resolution of
discontinuities.
, ,100871
E-mail: hztang@pku.edu.cn
An Introduction to Piecewise Algebraic Variety
, ,116024
E-mail: renhong@dlut.edu.cn
It is well known that the face of any object in the nature is combined by some sur-
faces, and their intersection. Therefore to study the geometric properties on intersection
of surfaces will be very important in both theory and practice.
The algebraic variety is one of most important subjects in the classical algebraic ge-
ometry which deals with the geometric properties of intersection of surfaces defined by
multivariate polynomials.
Because of most surfaces have been recently represented by multivariate piecewise
polynomials, the study of so-called “piecewise algebraic variety” defined as an intersec-
tion of surfaces represented by multivariate piecewise polynomials should be also most
important in both theory and practice. In fact, the piecewise algebraic variety is also
very useful in Computer Aided Design, Computer Aided Geometric Design, Computer
Aided Manufacture, Computer Aided Engineering, Computer Graphics, as well as Image
Processing.
The main problem is how we can study the piecewise algebraic variety. However,
there are some essential difficulties in studying the piecewise algebraic variety, because
38. 33
of some reasons as follows:
• The multivariate piecewise polynomial has very strong local property;
• Zeros of a multivariate polynomial determined by multivariate piecewise polynomial
on a given cell may be outside this cell;
• The multivariate piecewise polynomial defined on a given partition of a domain not
only depends on topological property of the partition, but also sometimes depends on
geometric property of the partition.
In fact, most methods for studying the classical algebraic variety can not be used to
study the piecewise algebraic variety. Therefore, we have to find some special methods
to study the piecewise algebraic variety, much more to study “real piece-wise algebraic
variety”.
In this talk, we will introduce some results on the piecewise algebraic variety. Moreover
some open problems will also be introduced too.
, , 100088
E-mail: wang shuanghu@iapcm.ac.cn
jet
, ,116023
E-mail: wuweiw@dlut.edu.cn
An Adaptive Edge Element Method with Perfectly Matched
Absorbing Layers for Wave Scattering by Bipe
39. 34
, ,210093
E-mail: hjw@nju.edu.cn
An edge element adaptive strategy with error control is developed for wave scattering
by biperiodic structures. The unbounded computational domain is truncated to a bounded
one by a perfectly matched layer (PML) technique. The PML parameters, such as the
thickness of the layer and the medium properties, are determined through sharp a poste-
riori error estimates. Numerical experiments are presented to illustrate the competitive
behavior of the proposed adaptive method.
, ,361005
E-mail: cjxu@xmu.edu.cn
In this talk, we present some new developments in spectral element methods (SEM) for
the unsteady Navier-Stokes equations. The main ingredients include:
1) a fast PN × PN spectral element solver using preconditioned Schur complement al-
gorithm for the Navier-Stokes equations, and detailed comparison and discussion of some
new approaches. The link of different methods will be clarified. The key feature of our
method is that only one grid is needed for the velocity and pressure approximations. Al-
though not yet proven by rigorous theoretical analysis, the stability and accuracy of this
simple method are demonstrated by a series of the numerical experiment.
2) an efficient stabilization method, which consists in employing SVV technique in
the standard SEM for accurate computations of high-Reynolds number flows. Our new
formulation yields an algorithm which can be easily implemented and does not require
additional computational time.
3) applications of the Legendre-SVV-SEM to the LES of turbulent flows. The SVV-
SEM is used as a no-model approach, i.e. no modeling of the sub-grid scale tensor which
results from the spatial filtering of the Navier-Stokes equations. Some simulation in the
3D driven cavity flow and the wake flow behind a circular cylinder give very satisfactory
results.
4) development and implementation of a triangular spectral element method for the
Navier-Stokes equations in complex geometries.
40. 35
Relations Between the Multiscale Methods for Elliptic
Homogenization Problems
, ,230026
E-mail: xyyue@ustc.edu.cn
We will give a short review on multiscale methods for elliptic homogenization prob-
lems. We will focus on the intrinsic links between some popular methods such as gener-
alized finite element methods (GFEM), residual- free bubble methods (RFB), variational
multiscale methods (VMS), multiscale finite element method (MSFEM) and heteroge-
neous multiscale methods (HMM).
Numerical Investigation for a Macroscopic Model for
Hydrodynamic Nematic Liquid Crystals
, , 100875
E-mail: hzhang@bnu.edu.cn
We use finite element methods to simulate the hydrodynamical systems governing the
motions of nematic liquid crystals in a bounded domain. We reformulate the original
model in the weak form which is consistent with the continuous dissipative energy law
for the flow and director fields. This enables us to use convenient conformal C 0 finite
elements in solving the problem. Moreover, a discrete energy law is derived for a modified
midpoint time discretization scheme. A fixed iterative method is used to solve the resulted
nonlinear system so that a matrix free time evolution may be achieved and velocity and
director variables may be solved separately. A number of hydrodynamical liquid crystal
examples are computed to demonstrate the effects of the parameters and the performance
of the method.
At last we will present an example of this macroscopic model for complex fluids in
“1+2” dimension case. It be found that the direction of the molecules will tumble from
the boundary layer and later on the inner layer with a much longer time period. This is
consistent with the theoretical predict of special case. Moreover, we find some complex
phenomena, where the tumbling rises from boundary layer then is deep into the middle
41. 36
area more clearly when the viscosity coefficient of the macro flow has a larger value. The
norm of the molecular director would endure greater change as well. This implies that the
viscosity of flow plays the role of an accelerator in the whole complex fluids. Comparing
these results with the theoretical analysis, we can find that the gradient of the velocity has
direct impact on the tumbling phenomena. These results roughly show that such a scheme
is capable of giving rich phenomena embedded in the macro-micro model.
42. 37
E
Design of Computer Experiments: Construction and Theory
, ,100871
E-mail: myai@math.pku.edu.cn
Computer models are widely used in business, engineering, and sciences to study com-
plex real world systems. The corresponding physical experimentation might otherwise be
time-consuming, costly, or even infeasible to conduct. Space-filling designs have been
widely used for conducting computer experiments. A large computer code, like a finite
element analysis model, is often run at variable degrees of sophistication, resulting in mul-
tiple computer experiments with different levels of accuracy and varying computational
times. In this talk we consider the situation in which two such experiments are available
and one source is generally more accurate than the other but also more expensive to run.
A new type of designs called nested space-filling designs has been proposed for conduct-
ing multiple computer experiments with different levels of accuracy. Several approaches
to constructing such designs have also been developed. The development of these meth-
ods also leads to the introduction of several new discrete mathematics concepts, including
nested orthogonal arrays and nested difference matrices.
Feynman-Kac
, ,571158
E-mail: czchen@hainnu.edu.cn
Feynman-Kac
(E, D(E)) Xt (E, D(E)) u ∈ D(E) u(Xt )
Fukushima
u(Xt ) − u(X0 ) = Mtu + Ntu
Mtu Ntu Fukushima
Ntu
43. 38
u
Pu f (x) = E x [e−Nt f (Xt )]
t
Feynman-Kac Feynman-Kac
Pu Ntu
Feynman-Kac t
Girsanov h−
Pu
t
Homeomorphic Flows and Large Deviations for Stochastic
Differential Equations Driven by a G-Brownian Motion
, ,430072
E-mail: fqgao@whu.edu.cn
Some moment estimates and H¨ lder continuity of the G-stochastic integral and the
o
solutions of stochastic differential equations driven by a G-Brownian motion are obtained.
Homeomorphic property with respect to the initial values and large deviation principle for
the stochastic differential equations are established.
Markov Skeleton Processes and Their Applications
, ,410075
E-mail: zthou@csu.edu.cn
A stochastic process X(t) is called a Markov skeleton process(MSP) if it has the Markov
property on a sequence of stopping times τn ↑ ∞. The usual Markov process, semi-
Markov process, deterministic Markov process and semi-regenerative process can be re-
garded as special cases of MSP. In this paper, first, backward and forward equations with
which we can compute one-dimensional distribution is derived, and then formulas to com-
pute finite-dimensional distribution and the existence and computation of limit distribu-
tion are also obtained. Based ourselves upon the above the results, we give a tentative
study of queueing system, reliability system, and storage system. Transient distribu-
tion and formulas to compute limit distribution of the stochastic processes introduced
for studying these system are presented in the latter half the paper.
44. 39
Fisher :
, ,100190
E-mail: luosl@amt.ac.cn
Fisher
Fisher
, ,510275
E-mail: renjg@mail.sysu.edu.cn
Limit Theorems for Super-diffusions and Branching Hunt Processes
, ,100871
E-mail: yxren@math.pku.edu.cn
First, we establish a scaling limit theorem for a large class of Dawson-Watanabe su-
perprocesses whose underlying spatial motions are symmetric Hunt processes, where the
convergence is in the sense of convergence in probability. When the underling process is a
symmetric diffusion with Cb -coefficients or a symmetric Levy process on Rd whose Levy
1
exponent Ψ(η) is bounded from below by c|η|α for some c > 0 and α ∈ (0, 2) when |η| is
large, a stronger almost sure limit theorem is established for the superprocess. Second, we
establish Kesten-Stigum L log L type theorems for super-diffusions and branching Hunt
processes using spine decompositions.
Wiener
45. 40
, ,100875
E-mail: shaojh@bnu.edu.cn
Wiener
Ambrosio-Gigli-Savare
Ornstein-Uhlenbeck
Additive Hazards Regression with Censoring Indicators
missing at random
, ,100190
E-mail: slq@amt.ac.cn
In this article, we consider a semiparametric additive hazards regression model for
right-censored data that allows some censoring indicators to be missing at random. We
develop a class of estimating equations and use an inverse probability weighted approach
to estimate the regression parameters. Nonparametric smoothing techniques are employed
to estimate the probability of non-missingness and the conditional probability of an un-
censored observation. The asymptotic properties of the resulting estimators are derived.
Simulation studies show that the proposed estimators perform well. We motivate and
illustrate our methods with data from a brain cancer clinical trial.
Model Selection Criteria for Missing Data Problems
via the EM Algorithm
, ,650091
E-mail: nstang@ynu.edu.cn
We propose a class of novel and computationally attractive model selection criteria
for missing data problems based on the output of the expectation-maximization (EM)
algorithm. The proposed criteria do not require numerical integration or Laplace approx-
imations to integrals. The methodology is very general and can be applied to numerous
46. 41
situations involving incomplete data within an EM framework, from covariates missing
at random (MAR) in arbitrary regression models to nonignorably missing longitudinal
responses and/or covariates. Towards this goal, we develop a class of information criteria
for missing data problems, called ICH, Q, which yields the Akaike information criterion
(AIC) and the Bayesian Information Criterion (BIC) as special cases. The computation
of ICH, Q requires an analytic approximation to a complicated function, called the H-
function, along with output from the EM algorithm used in obtaining maximum likelihood
estimates. To eliminate the analytic approximation to the H-function, we also propose a
computationally simpler approximation to ICH, Q, called ICQ, whose computation de-
pends solely on the Q-function of the EM algorithm. Advantages and disadvantages of
ICH, Q and ICQ are discussed and examined in detail in the context of missing data prob-
lems. Theoretical properties of ICH, Q and ICQ are also investigated in detail, including
consistency. Extensive simulations are given to demonstrate the methodology and exam-
ine the small and large sample performance of ICH, Q and ICQ in missing data problems.
An AIDS data set is also presented to illustrate the proposed methodology.
profile
, ,300071
E-mail: zjwang@nankai.edu.cn
linear profile, general profile, nonparametric profile
Stochastic Maximum Principles for Partially Observed Optimal
Control Problems and Applications
, ,250100
E-mail: wuzhen@sdu.edu.cn
In this talk, we first give some stochastic maximum principles-the necessary condition
of the optimal control–for partially observed risk-sensitive optimal control problems. Ap-
plying these theoretical results we study a partially observed linear-quadratic non-zero
sum stochastic differential game problem and give an explicit observable Nash equilib-
rium point. As a natural deduction, a general maximum principle is also given for a fully
47. 42
observed risk-sensitive case. This result is applied to study a risk-sensitive optimal portfo-
lio problem. An explicit optimal investment strategy and a cost functional are obtained. A
numerical simulation result shows an influence of a risk-sensitive parameter on an optimal
investment proportion, this coincides with its economic meaning and theoretical results.
At last, we study a partially observed stochastic recursive optimal control problems. This
kind of problems have wide applications in finance and economic such as the celebrated
principal-agent problems. We also give a maximum principle for this kind of optimization
problems in this talk.
, ,100081
E-mail: xuxz@bit.edu.cnn
p
p
Fiducial
p
p
Fiducial
Estimation for a Partial-linear Single-index Model
, ,100124
E-mail: lgxue@bjut.edu.cn
In this paper, we study the estimation for a partial-linear single-index model. A two-
stage estimation procedure is proposed to estimate the link function for the single index
and the parameters in the single index, as well as the parameters in the linear component
of the model. Asymptotic normality is established for both parametric components. For
the index, a constrained estimating equation leads to an asymptotically more efficient
48. 43
estimator than existing estimators in the sense that it is of a smaller limiting variance.
The estimator of the nonparametric link function achieves optimal convergence rates; and
the structural error variance is obtained. In addition, the results facilitate the construction
of confidence regions and hypothesis testing for the unknown parameters. A simulation
study is performed and an application to a real dataset is illustrated. The extension to
multiple indices is briefly sketched.
Markov
, ,200433
E-mail: jgying@fudan.edu.cn
In this paper, we shall prove that the irreducibility in the sense of fine topology implies
the uniqueness of invariant probability measures. It is also proven that thisirreducibility is
strictly weaker than the strong Feller property plus irreducibility in the sense of original
topology, which is the usual uniqueness condition.
Generalized Profile LSE in Varying-Coefficient Partially Linear
Models with Measurement Errors
/ , / ,100190/200433
E-mail: yzhou@amss.ac.cn
This paper is concerned with the estimating problem of a semi-parametric varying-
coefficient partially linear errors-in-variables model. Due to the measurement errors the
usual profile least squares estimator of the parametric component, local polynomial es-
timator of the nonparametric component and profile least squares based estimator of the
error variance are biased and inconsistent. By taking the measurement errors into account
we propose a generalized profile least squares estimator for the parametric component
and show it is consistent and asymptotically normal. Correspondingly, the consistent esti-
mations of the nonparametric component and error variance are proposed as well. These
results can be used to make asymptotically valid statistical inferences. Some simulation
studies are conducted to illustrate the finite sample performances of these proposed esti-
mations.
50. 45
F
Primal-Dual IPMs for Conic Optimization
, ,200444
E-mail: yqbai@shu.edu.cn
After the path-breaking paper of Karmarkar, the Interior-point Methods (IPMs) has
been among the most effective methods for solving linear optimization problems. With
the development of IPMs, it has been extended to the conic optimization, a class of special
convex optimization such as second-order cone optimization and semidefinite optimiza-
tion. Because the scheme of primal-dual IPMs includes three blocks: the selection of bar-
rier functions, the method of unconstrained optimization and the policy of updating the
penalty parameter. We introduce a new class of univariate function called kernel function
which is defied by some simple conditions. The barrier function is determined by kernel
function with nice properties. These properties enable us to derive many new and tight
estimates that greatly simplify the analysis of IPMs based on the kernel functions. Both
in the algorithm and in its analysis we use a single neighborhood of the central path; the
neighborhood naturally depends on the kernel function. An important conclusion is that
inverse functions of suitable restrictions of the kernel function and its first derivative more
or less determine the behavior of the corresponding IPMs. Based on the new estimates we
present a simple and unified computational scheme for the complexity analysis of primal-
dual IPMs for conic optimization, including linear optimization, second-order cone opti-
mization and semidefinite optimization. Iteration bounds both for large- and small-update
methods are derived. It is shown that small-update methods based on the new kernel func-
√
tions all have the same complexity as the classical primal-dual IPM, namely O( nlog n ).
√
For large-update methods the best obtained bound is O( n(logn)log n ), which is up till
now the best known bound for such methods.
, ,100190
E-mail: htfang@iss.ac.cn
51. 46
, ,
Wiener
Applications of Steiner Tree Problem in Network Design
, ,100190
E-mail: xdhu@amss.ac.cn
The study of the Steiner tree problem has received lots of attentions since many impor-
tant open problems were solved in 1990s. Those achievements have greatly influenced not
only the general theory of designs and analysis of approximation algorithms for combina-
torial optimizations but also the discovery and research on many new important applica-
tions. Those applications usually require some modifications on the classical Steiner tree
problem, and hence they demand new techniques for solving them. As a result, studying
various Steiner tree problems forms a very hot topic in the past decade.In this talk we will
study some of those applications to the design of computer communication networks.
Error estimates for an optimal control problem
of the heat equation with state and control constrain
, ,430072
E-mail: wanggs62@yeah.net
We study the priori error analysis for numerical approximations of the optimal con-
trol governed by the heat equation with certain control constraint and ending point state
constraint. By making use of the classical space-time discretization scheme, namely,
finite element method with the space variable and implicit Euler discretization for the
time variable, We first project the original optimal control problem into a semi-discrete
52. 47
control and state constrained optimal control problem governed by an ordinary differ-
ential equation, and then project the aforementioned semi-discrete problem into a fully
discrete optimization problem with constraints. With the help of Pontryagin maximum
principle, we obtain, under a certain reasonable condition of Slater style, not only a con-
vergence order for L2 −error of the optimal controls between the original problem and
the semi-discrete problems, but also the solutions between the semi-discrete problem and
fully discrete problem. Thus, we provide a base to compute the optimal control of the
heat equation with certain control constraint and ending point state constraint. The main
difficulty arising in this work is the lack of the quantitative analysis for the Lagrange mul-
tipliers appeared in Pontryagin maximum principles in the original problem and discrete
problems. These multipliers are produced by the state constraint.
, ,100081
E-mail: jmwang@bit.edu.cn
Riesz
Riesz 1
2
3
Riesz
Riesz
, ,100190
E-mail: zrxi@iss.ac.cn
53. 48
I
A
B ,
A
C
II Markov
10-6
Markov
Markov
Markov
Markov Markov
III
54. 49
A. 0
B. 0
Lowner
, ,100044
E-mail: nhxiu@bjtu.edu.cn
1934 Karl Lowner Sun-Sun
Lowner
Lowner Lowner-Heinz
Lowner-type Taylor
1 Lowner Euclidean Jordan Algebra 2 Lowner
3 Lowner 4
Inverse and Generalized Inverse Optimization Problems
, ,100190
E-mail: xgyang@iss.ac.cn
The inverse optimization problems are to find a minimal modification cost of the param-
eters of the original problem such that some given solutions become optimum under the
new parameters, while the generalized inverse optimization problems are to find a mini-
mal modification cost of the parameters of the original problem such that a given objective
is achieved. In this talk, we will give a survey about the applications, major developments,
major techniques of the inverse and generalized inverse optimization problems. Then we
present a unified model which can combine the inverse and generalized inverse problems.