µ±Ç°Î»ÖãºÊ×Ò³ > ѧÊõ±¨¸æ > 2009ÄêѧÊõ±¨¸æ

±¨¸æÌâÄ¿£º ÊýѧÔÚ¿ó²ØÑо¿ÖеÄÓ¦ÓÃ
±¨ ¸æ ÈË£º Prof. Fun-Den Wang, Colorado School of Mines, USA
ʱ¼äµØµã£º 2009Äê11ÔÂ11ÈÕÉÏÎç9:00 S1013
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º On a degenerate free-boundary value problem arising from subsonic-sonic flows(ϵÁб¨¸æ04)
±¨ ¸æ ÈË£º ÐÁÖÜÆ½½ÌÊÚ (Ïã¸ÛÖÐÎÄ´óѧ§)
ʱ¼äµØµã£º 2009Äê7ÔÂ24ÈÕÉÏÎç10:30 ³¿ÐËÖÐÐÄ510ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º Navier-Stokes·½³Ì±ß½çÐÎ×´¿ØÖƺÍάÊý·ÖÁÑËã·¨
±¨ ¸æ ÈË£º »Æ°¬Ïã½ÌÊÚ (Î÷°²½»Í¨´óѧ)
ʱ¼äµØµã£º 2009Äê7ÔÂ13ÈÕÏÂÎç4:00 ˼Դ¥1013ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º Levenberg-Marquardt method and trust region method for nonlinear equations
±¨ ¸æ ÈË£º ·¶½ðÑà ½ÌÊÚ (ÉϺ£½»Í¨´óѧ)
ʱ¼äµØµã£º 2009Äê7ÔÂ13ÈÕÏÂÎç3:50 ˼Դ¥703ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º Warehouse-Retailer Network Design Problem with Submodular Penalties: Heuristics and Bounds
±¨ ¸æ ÈË£º Ðì´ó´¨ ½ÌÊÚ (±±¾©¹¤Òµ´óѧ)
ʱ¼äµØµã£º 2009Äê7ÔÂ13ÈÕÏÂÎç2:50 ˼Դ¥703ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º Semismooth Newton methods for SVM
±¨ ¸æ ÈË£º Prof.Hongxia Yin (Minnesota State University)
ʱ¼äµØµã£º 2009Äê7ÔÂ13ÈÕÏÂÎç2:00 ˼Դ¥703ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º The report on basic scientific directions of Institute for system dynamics and control theory
±¨ ¸æ ÈË£º Igor BychkovԺʿ (¶íÂÞ˹¿ÆÑ§ÔºÎ÷²®ÀûÑÇ·ÖԺϵͳ¶¯Á¦Ñ§Óë¿ØÖÆÀíÂÛÑо¿ËùËù³¤)
ʱ¼äµØµã£º 2009Äê6ÔÂ23ÈÕÉÏÎç10:00 ³¿ÐË110ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º Some Recent Progress on Switching Diffusion Systems (ϵÁб¨¸æ03)
±¨ ¸æ ÈË£º Prof. G. Yin (Department of Mathematics Wayne State University)
ʱ¼äµØµã£º 2009Äê6ÔÂ22ÈÕÏÂÎç2:30 ˼Դ¥1013ÊÒ
Õª       Òª£º In this talk, we report some of our recent work on switching diffusion processes in which continuous dynamics and discrete events coexist. With motivation from treating control and optimization of stochastic systems, we study stability problems. After recalling the notion of regularity and recurrence, necessary and sufficient conditions for positive recurrence will be provided; ergodicity will be examined; stability about an equilibrium point will be studied; rates of convergence of numerical algorithms for optimal control of switching diffusions will be mentioned.
 
±¨¸æÌâÄ¿£º Nonlinear stability of traveling waves from conservation laws with higher order diffusion
±¨ ¸æ ÈË£º Prof. Changbing Hu (University of Louisville, USA)
ʱ¼äµØµã£º 2009Äê6ÔÂ18ÈÕÏÂÎç4:00 ˼Դ¥703ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º Special geometry and holomorphic anomaly Equations for generalized Calabi-Yau manifolds
±¨ ¸æ ÈË£º ºúÉ­ ½ÌÊÚ (Öйú¿ÆÑ§¼¼Êõ´óѧÊýѧϵ)
ʱ¼äµØµã£º 2009Äê6ÔÂ17ÈÕÏÂÎç4:00 ³¿ÐË410ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º ÕýÇúÂÊÁ÷ÐεķÖÀàÎÊÌâ
±¨ ¸æ ÈË£º Öìì䯽 ½ÌÊÚ (ÖÐɽ´óѧ)
ʱ¼äµØµã£º 2009Äê6ÔÂ16ÈÕÉÏÎç10:00 ˼Դ¥1013ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º ÕýÇúÂÊÁ÷ÐεķÖÀàÎÊÌâ
±¨ ¸æ ÈË£º Öìì䯽 ½ÌÊÚ (ÖÐɽ´óѧ)
ʱ¼äµØµã£º 2009Äê6ÔÂ16ÈÕÉÏÎç10:00 ˼Դ¥1013ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º On the uniqueness of invariant measure for stochastic delay equations
±¨ ¸æ ÈË£º Prof. Michael Scheutzow (Technical University of Berlin, Germany)
ʱ¼äµØµã£º 2009Äê6ÔÂ15ÈÕÏÂÎç4:00 ˼Դ¥1013ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º Three Types of Exact Tail Asymptotics for Random Walks in the Quarter Plane
±¨ ¸æ ÈË£º Prof. Tao Luo (Georgetown University, USA)
ʱ¼äµØµã£º 2009Äê6ÔÂ11ÈÕÏÂÎç4:00 ˼Դ¥712ÊÒ
Õª       Òª£º In this talk, I will present some results on the stability of rotating and non-rotating white dwarfs stars for the compressible Euler-Poisson equations, if the total mass is less than a critical mass. This is a joint work with Joel Smoller.
 
±¨¸æÌâÄ¿£º Three Types of Exact Tail Asymptotics for Random Walks in the Quarter Plane
±¨ ¸æ ÈË£º Dr. Yiqiang Q. Zhao (Carleton University, Canada)
ʱ¼äµØµã£º 2009Äê6ÔÂ10ÈÕÉÏÎç10:30 ˼Դ¥712ÊÒ
Õª       Òª£º Random walks in the quarter plane can be viewed as a discrete time Markov chain with a two-dimensional state space. Under the stability condition, examples demonstrated that for a fixed value of one variable, the stationary tail probability as the other variable goes large decays according to three types: (a) exact geometric; (b) geometric with the subgeometric factor $n^{-3/2}$; and (c) geometric with the subgeometric factor $n^{-1/2}$. It is believed that these are the only types of tail asymptotics for any random walk in the quarter plane. Under some irreducible (or non-singular) conditions, this has been almost proved. In this talk, we discuss (i) why only these three types of tail asymptotics are expected; and (ii) other types may exist if the irreducible (or non-singular) condition does not hold. We also ask the question: are these three the only types of exact tail asymptotics for more general random walks, say random walks with jumps, assuming suitable irreducibility?
 
±¨¸æÌâÄ¿£º ¶àµã±ßÖµÎÊÌâµÄ¶à¸ö±äºÅ½â
±¨ ¸æ ÈË£º ÐìÎ÷°² ½ÌÊÚ (ÐìÖÝʦ·¶´óѧ)
ʱ¼äµØµã£º 2009Äê5ÔÂ27ÈÕÉÏÎç11:00 ˼Դ¥712ÊÒ
Õª       Òª£º ¶àµã±ßÖµÎÊÌâÊǽüÄê·ÇÏßÐÔ³£Î¢·Ö·½³ÌÑо¿ÁìÓòµÄÈȵãÎÊÌâÖ®Ò»£¬ÆäÌØÊâÇéÐΰüº¬Á˾­µäµÄÁ½µã±ßÖµÎÊÌâ¡£½üÄêÀ´ÈËÃÇ½Ï¶àµØÑо¿Á˶àµã±ßÖµÎÊÌâÕý½âºÍ·Çƽ·²½âµÄ´æÔÚÎÊÌâ¡£Ñо¿¶àµã±ßÖµÎÊÌâ±äºÅ½âµÄ´æÔÚÐÔÎÊÌâÊǸöÓÐÒâÒåµÄ¹¤×÷£¬´æÔÚһЩеÄÀ§ÄÑ¡£ ±¨¸æÈ˽«½éÉÜ»ùÓÚÍØÆË¶ÈºÍÈ«¾Ö½á¹¹ÀíÂÛ£¬ÌرðÊÇÓëÉÏ¡¢Ï½â½áºÏѰÕÒ¶àµã±ßÖµÎÊÌâ¶à¸ö±äºÅ½âµÄһЩз½·¨¡£
 
±¨¸æÌâÄ¿£º Convergence properties of solutions to several classes PDEs
±¨ ¸æ ÈË£º Prof. Lei Zhang (University of Alabama at Birmingham£USA)
ʱ¼äµØµã£º 2009Äê5ÔÂ27ÈÕÉÏÎç10:00 ˼Դ¥712ÊÒ
Õª       Òª£º Many problems in Geometry, Physics and Biology are described by nonlinear partial differential equations of second order or four order. In this talk I shall mainly mention a class of fourth order equations from conformal geometry and some systems of second order equations with rich background in Physics and Ecology. These equations are challenging in analysis and represent some typical difficulties. There are some important open problems related to these equations and I will report the recent progress on these problems.
 
±¨¸æÌâÄ¿£º Strong maximum principle for singular solutions of nonlinear elliptic equations
±¨ ¸æ ÈË£º Professor Yanyan Li (Rutgers University, USA)
ʱ¼äµØµã£º 2009Äê5ÔÂ25ÈÕÉÏÎç10:00 ˼Դ¥1013ÊÒ
Õª       Òª£º  
 
±¨¸æÌâÄ¿£º Transformation Optics and Three Dimensional Cloaking of Active Devices
±¨ ¸æ ÈË£º Professor Hongyu Liu (Dept of Mathematics, University of Washington, USA)
ʱ¼äµØµã£º 2009Äê5ÔÂ22ÈÕÏÂÎç4:00 ˼Դ¥1013ÊÒ
Õª       Òª£º .In this talk, we shall first discuss the transformation optics in the design of cloaking devices to obtain customized electromagnetic wave propogation. Then, using the transformation optics, we give the construction of a three dimensional cloaking device, which makes the cloaked region with a source invisible to wave detection. Based on a new notion of "finite energy solution" for singular PDE, the rigorous mathematical justification of the invisibility cloaking is given. Finally, we briefly introduce a specially designed finite element discretization for numerically solving the cloaking problem
 
±¨¸æÌâÄ¿£º Navier£­Stokes equations in the critical Lebesgue spaces
±¨ ¸æ ÈË£º ¶¼´óÅô²©Ê¿ (¸´µ©´óѧÊýѧ¿ÆÑ§Ñ§Ôº)
ʱ¼äµØµã£º 2009Äê5ÔÂ21ÈÕÏÂÎç4:00 ˼Դ¥1013ÊÒ
Õª       Òª£º .
 
±¨¸æÌâÄ¿£º Two-stage efficient estimation of stratified mark-specific proportional hazards models with missing marks£¨ÏµÁб¨¸æ2£©
±¨ ¸æ ÈË£º Professor Yanqing Sun (University of North Carolina at Charlotte)
ʱ¼äµØµã£º 2009Äê2ÔÂ10ÈÕÏÂÎç4:00 ˼Դ¥712ÊÒ
Õª       Òª£º Motivated by the need to assess HIV vaccine efficacy, the competing risks model with a continuum of competing risks, called mark, has been studied in recent years by a few authors. The mark-specific proportional hazards model is the proportional hazards model for mark-specific hazard function in which the regression parameters depend nonparametrically on the mark and the baseline hazard to depend nonparametrically on both time and mark. In the HIV vaccine efficacy trial, a very high percentage of marks of interest may be missing for infected trial participants due to the evolving nature of the HIV viruses. This research investigates the stratified mark-specific proportional hazards model with missing marks where the baseline functions may vary with strata. A two-stage estimation procedure is proposed. In the first stage, we propose an estimation procedure based on the inverse probability weighted complete-case technique. A more efficient augmented inverse probability weighted complete-case estimator is derived in the second stage based on the knowledge of the distribution of marks. In addition, the proposed two-stage procedure can utilize auxiliary information for missing marks to improve efficiency. The asymptotic properties of the proposed estimators are investigated. We show that the AUG estimator is more efficient than the IPW estimator. The finite sample performance of the proposed methods are evaluated through simulations. The applications to the HIV vaccine efficacy assessment are discussed .
 
±¨¸æÌâÄ¿£º Jackknife empirical likelihood£¨ÏµÁб¨¸æ1£©
±¨ ¸æ ÈË£º Professor Bing-Yi JING (Hong Kong Univ. of Science & Technology)
ʱ¼äµØµã£º 2009Äê1ÔÂ7ÈÕÏÂÎç4:00 ˼Դ¥703ÊÒ
Õª       Òª£º Empirical likelihood, since its first introduction by Owen in 1988, has been found very useful in many different occasions. However, it is not always trouble-free. For instance, when applied directly to some more complicated statistics such as U-statistics, it runs into serious computational difficulties, particularly when the sample size gets larger. In this paper, we will take a fresh look at the problem. We introduce a new approach which combines two of popular nonparametric approaches: the jackknife and the empirical likelihood. The resulting method, referred to as the jackknife empirical likelihood (JEL), is extremely simple to use in practice, in comparison with other existing methods. In particular, the new method is shown to be very effictive in handling U-statistics. The JEL can be potentially useful for other nonlinear statistics, however, further research is needed to justify its usage in those situations.
(This is joint work with Junqing YUAN and Want ZHOU.) .
 
±¨¸æÌâÄ¿£º Quasi-Likelihood of Models: Modified Profile Likelihood for Model Selection
±¨ ¸æ ÈË£º Dr. He Heping (The University of Kansas )
ʱ¼äµØµã£º 2009Äê1ÔÂ5ÈÕÏÂÎç3:00 ˼Դ¥703ÊÒ
Õª       Òª£º The goal of this research is to develop the modified profile likelihoods (MPLs) as quasi-likelihood functions of models for model selection. It treats those parameters in each possible model as nuisance parameters, and the models themselves as the values of the ``parameter'' of interest, then uses the likelihood ratio approximation to conditional distribution of maximum likelihood estimators of the parameters in the possible models to develop the MPLs of models. Note that the MPL proposed here is considered as a function of models themselves instead of parameters in the models. The selected model is then the one maximizing the quasi-likelihood of models, i.e. the MPL. Some simulations have show that the MPL works very well for the selection of error probability laws in location-scale models. The MPL of models has also been obtained for composite transformation models. In future, this research may develop the MPL of models for model selection problems in exponential family, regular models, regressions, change point problems, AR or ARMA time series models and so on.
 
µ±Ç°Î»ÖãºÊ×Ò³ > ѧÊõ±¨¸æ > 2009ÄêѧÊõ±¨¸æ