with Perfect Discrete Time Observations Campus de Beaulieu for all z 2 R d M z = h?1 (z) = fx 2 R m : h(x) = zg :
|
|
- Eleanore Martin
- 5 years ago
- Views:
Transcription
1 Nonlinear Filtering with Perfect Discrete Time Observations Marc Joannides Department of Mathematics Imperial College London SW7 2B, United Kingdom Francois LeGland IRISA / INRIA Campus de Beaulieu 3502 Rennes Cedex, France legland@irisa.fr Abstract : We consider the problem of estimating the state of a diusion process, based on discrete time observations in singular noise. We reduce the problem to a static problem, and we show that the solution is provided by the area or co{area formula of geometric measure theory, provided the observed value is a regular value of the observation function. In order to address the case of singular values, we propose another approach, based on small{noise perturbation and asymptotics of Laplace integrals. 1 INTRODUCTION One major limiting assumption in the nonlinear ltering literature is the non degeneracy of the observation noise covariance matrix. However, there are numerous situations of practical interest, where some perfect noise{free information is available about the unknown state, and yet the problem of state estimation with degenerate observation noise has received little attention, except in the linear case. An additional motivation for studying this problem, is the existing connection with various problems of state estimation for non classical dynamical systems, including : hybrid systems, i.e. systems with state constraints, stochastic dierential{algebraic systems, systems with colored noise, see Korezlioglu and Runggaldier [8], systems with state{dependent observation noise, see Takeuchi and Akashi [11], etc. Finally, a better understanding of the nonlinear ltering problem with noise{free observations should help designing robust and ecient numerical approximation schemes in the important case where the observation noise is small. To be more specic, we consider the following state equation in R m dx t b(x t ) dt + (X t ) dw t ; (1) where fw t ; t 0g is a Wiener process of appropriate dimension, and the noise{free d{dimensional discrete time observations z k h(x tk ) : This work was partially supported by the Army Research Oce, under grant DAAH0{95{1{016. The objective of nonlinear ltering is to compute the conditional probability distribution k (dx) P[X tk 2 dx j z 1 ; ; z k ] : What makes this problem singular is that the state X tk is known exactly to belong to the level set k, where for all z 2 R d h?1 (z) fx 2 R m : h(x) zg : Therefore, the conditional probability distribution k is supported by the set k, which in general has null Lebesgue measure, and k does not have a density w.r.t. the Lebesgue measure on R m. The question naturally arises whether k has a density w.r.t. some canonical measure on the level set k. The objective of this paper is to give conditions under which this holds, and to provide an explicit expression for the density. As usual, the transition from k to k+1 consists in two steps. In the prediction step, i.e. between times t k and t k+1, we solve the Fokker{Planck k L k t ; k t k k ; where L is the adjoint of the innitesimal generator L associated with equation (1). For all t k t t k+1 it holds k t (dx) P[X t 2 dx j z 1 ; ; z k ] : and in particular at nal time t k+1 we obtain the prior probability distribution? k+1 k t k+1 of the state X tk+1. In the correction step, i.e. at time t k+1, we have to combine the prior probability distribution? k+1 and the new observation z k+1 which denes the level set k+1, so as to obtain the posterior probability distribution k+1 of the state X tk+1. We make now the following two assumptions.
2 Assumption A : For all 0 s < t, and all x 2 R m, the conditional probability distribution s;x t (dy) P[X t 2 dy j X s x] ; which is the solution of the Fokker{Planck s;x L s;x t ; s;x s x ; has a continuous density w.r.t. the Lebesgue measure on R m. Assumption B : The mapping h is continuously dierentiable from R m to R d. The purpose of Assumption A is to make sure that, even though k is supported by a set of null Lebesgue measure, the prior probability distribution? has k+1 a continuous density w.r.t. the Lebesgue measure on R m. Assumption B is used to describe the level sets. A stronger assumption (Assumption C) will be introduced later. Under Assumptions A and B, the original problem reduces as follows : Let X be a r.v. in R m, with absolutely continuous probability distribution, and continuous density p w.r.t. the Lebesgue measure on R m. Let the mapping h be continuously dierentiable from R m to R d. Compute the conditional probability distribution z (dx) P[X 2 dx j h(x) z] ; for all z 2 R d. Our objective is to provide an explicit expression for the conditional probability distribution z. In particular, assuming that a canonical measure z exists on the level set, we want to give conditions under which z is absolutely continuous w.r.t. z, and to provide an explicit expression for the density. 2 DIRECT APPROACH FOR REGULAR VALUES We begin this section with some denitions, and properties of the level sets associated with the continuously dierentiable mapping h from R m to R d. For all z 2 R d, the level set is dened by h?1 (z) fx 2 R m : h(x) zg : Notice that for all x 2 R m, h 0 (x) is a linear mapping from R m to R d, i.e. a d m matrix. By denition, x 2 R m is a regular point i the matrix h 0 (x) has full rank min(m; d). We distinguish two cases : In the case m d, x 2 R m is a regular point i h 0 (x) has rank m, i.e. J h (x) q det? [h 0 (x)] h 0 (x) > 0 : In the case m d, x 2 R m is a regular point i h 0 (x) has rank d, i.e. J h (x) q det? h 0 (x) [h 0 (x)] > 0 : Remark 2.1 If the mapping h is only Lipschitz continuous, it follows from the theorem of Rademacher, see Federer [6] or Evans and Gariepy [5, Chapter 3], that h is dierentiable a.e., and the Jacobian J h (x) is dened as above for a.e. x 2 R m. By denition, z 2 R d is a regular value i all x 2 are regular points. We distinguish again the same two cases : In the case m d, if z 2 R d is a regular value, then is a collection of isolated points, i.e. a 0{ dimensional submanifold of R m, and the canonical measure z on is the counting measure. In the case m d, if z 2 R d is a regular value, then is a (m?d){dimensional submanifold of R m, and the canonical measure z on is the Lebesgue measure on, see Berger and Gostiaux [1, Chapter 6]. Under additional smoothness assumption on the mapping h, it follows from the theorem of Sard that the set of regular values has full Lebesgue measure on R d, see Berger and Gostiaux [1, Chapter ] or Sternberg [10]. We are therefore focusing our attention below on regular values z 2 R d. We rst recall the area and co{area formula, see Federer [6] or Evans and Gariepy [5, Chapter 3]. Theorem 2.2 (Area formula) Let the mapping h be Lipschitz continuous from R m to R d, and assume m d. Then for every g 2 L 1 (R m ) R m g(x) J h (x) dx R d X x2 g(x) H m (dz) ; where H m denotes the m{dimensional Hausdor measure on R d. Theorem 2.3 (Co{area formula) Let the mapping h be Lipschitz continuous from R m to R d, and assume m d. Then for every g 2 L 1 (R m ), the restriction of g to the level set is H m?d {integrable for almost all z 2 R d, and g(x) J h (x) dx R m R d g(x) H m?d (dx) dz ; where H m?d denotes the (m? d){dimensional Hausdor measure on R m.
3 Remark 2. For a regular value z 2 R d, the restriction to of the (m? d){dimensional Hausdor measure agrees with the canonical measure z on. Remark 2.5 If the mapping h is only locally Lipschitz continuous from R m to R d, the formulas still hold for every g 2 L 1 (R m ) vanishing outside a compact subset of R m. Remark 2.6 If the mapping h is proper (i.e. h?1 (K) R m is compact, whenever K R d is compact), then all the level sets are compact, and in addition the set of regular values is open in R d. Indeed, assume the contrary, i.e. assume that there exists a sequence fz n ; n 1g in R d, converging to the regular value z. By denition, there exists a sequence fx n ; n 1g in R m such that h(x n ) z n and J h (x n ) 0 for all n 1. The converging sequence fz n ; n 1g stays in a compact K R d, hence the sequence fx n ; n 1g stays in h?1 (K) R m, which is compact since h is proper. Any limit point x of the sequence fx n ; n 1g satis- es h(x ) z and J h (x ) 0, which contradicts the assumption that z is a regular value. This motivates the introduction of the following stronger assumption. Assumption C : The mapping h is continuously dierentiable and proper from R m to R d. Under Assumption C, we deduce from the area and co{area formula, the following characterization of the conditional probability distribution z for a regular value z 2 R d. Theorem 2.7 Let Assumption C hold, and assume m d. Let z 2 R d be a regular value, with nite level set fx i ; i 2 Ig. The following summability condition holds X p(x i ) J h (x i ) < 1 : i2i If in addition the density p does not identically vanish on, then with density z X i2i q i x i ; p(x i ) J q i h (x i ) X : p(x j ) J j2i h (x j ) Theorem 2.8 Let Assumption C hold, and assume m d. Let z 2 R d be a regular value, and let z denote the canonical measure on the compact level set. The following integrability condition holds J h (x) z(dx) < 1 : If in addition the density p does not identically vanish on, then z (dx) q(x) z (dx) ; with density q(x) J h (x) : p(y) J h (y) z(dy) Proof. Under Assumption C, there exists a compact neighbourhood V z of the regular value z 2 R d, such that all values in V z are regular. For any test function dened in R m, and any Borel set B V z in R d, dene g(x) (x) 1 (h(x) 2 B) J h (x) : This mapping is Lebesgue{integrable and vanishes outside h?1 (V z ), which is a compact subset of R m. Therefore, it follows from Theorem 2.3 and Remark 2.5 that E[(X) 1 (h(x) 2 B) ] R m g(x) J h (x) dx R d 0 B 0 B 0 By taking (x) 1, we get P[h(X) 2 B] g(x) H m?d (dx) dz 0 (x) J h (x) Hm?d (dx) dz 0 (x) J h (x) z 0(dx) dz0 : B 0 J h (x) z 0(dx) dz0 : The result follows from the denition of E[(X) j h(x) z], see e.g. Breiman [2]. 2 3 ASYMPTOTICS OF LAPLACE INTEGRALS To obtain an expression of the conditional probability distribution z for any (regular or singular) value z 2 R d, and motivated by the design of robust and ef- cient numerical approximation schemes in nonlinear ltering with small observation noise, we introduce the following perturbation procedure : For " > 0, dene " + V " h(x) + V " ; where V " is a d{dimensional Gaussian r.v. with covariance matrix " I d, independent of the r.v. X. Using the Bayes formula, it is easy to dene a regular conditional probability distribution for the r.v. X given " : For any test function dened in R m E[(X) j " ] R m (x) exp? 1 2" 2 j"? h(x)j 2 dx exp? 1 R m 2" 2 j"? h(x)j 2 dx Two questions arise at this point : :
4 Does the left{hand side have a limit as " # 0, and does this limit provide a version of the conditional probability distribution of the r.v. X given, i.e. given h(x)? How can we compute the limit of the right{hand side as " # 0? The answer to the rst question is provided by the following result. Proposition 3.1 For any test function dened in R m E[ E[(X) j " ] j ]?! E[(X) j ] ; in L 1, as " # 0. Proof. Given a d{dimensional Gaussian white noise sequence fv n ; n 0g with unit covariance matrix, we dene e n + 1X kn k V k + e Vn ; with k "k? " k+1, and f" n ; n 0g is a sequence of positive numbers decreasing to zero. It follows from Takeuchi and Akashi [11, Lemma A.1] that E[(X) j e n ]?! E[(X) j ] ; a.s. and in L 1, as n! 0. As a result E[ E[(X) j e n ] j ]?! E[(X) j ] ; in L 1, as n! 0. Notice that (X; e n ) and (X; "n ) are identically distributed, hence E[ E[(X) j e n ] j ] E[ E[(X) j "n ] j ] a.s., and and therefore E[ E[(X) j "n ] j ]?! E[(X) j ] ; in L 1, as n! 0. Since the convergence holds for an arbitrary sequence f" n ; n 0g of positive numbers decreasing to zero, the result is proved. 2 Corollary 3.2 For any test function dened in R m R d " z? E[(X) j h(x) z] (dz)?! 0 ; as " # 0, where " z E[ E[(X) j " ] j h(x) z] ; and where denotes the probability distribution of the r.v. h(x). In the remaining of this section, we are going to answer the second question, i.e. to study the asymptotic behaviour of " z as " # 0, for a given z 2 Rd. For every x 2 R m, v 2 R d, we dene? " z(x; v) exp? 1 " 2 H" z(x; v) ; with and we notice that " z H " z (x; v) 1 2 jh(x)? (z + " v)j2 ; R d (x)? " R m z (x; v) dx R m? " z (x; v) dx e? 1 2 jvj2 dv (2) d2 : The rst step is to show that the behaviour of " z is determined only by the points x 2 R m in a neighbourhood of the level set. For this purpose, we introduce the following identiability condition. Assumption D : For all z 2 R d, r > 0 where M r z g z (r) inf x62m r z jh(x)? zj > 0 ; fx 2 R m : d(x; ) < rg : Remark 3.3 Under Assumption C, this identiability condition is always satised. Indeed, assume the contrary, i.e. assume that g z (r) 0 for some z 2 R d, r > 0. By denition, there exists a sequence fx n ; n 1g in R m nm r z such that h(x n) converges to z as n! 1. The converging sequence fh(x n ) ; n 1g stays in a compact K R d, hence the sequence fx n ; n 1g stays in h?1 (K) R m, which is compact since h is proper. Any limit point x of the sequence fx n ; n 1g satises h(x ) z and d(x ; ) r > 0, which is a contradiction. We dene D z and we notice that " z with D z In addition fv 2 R d : j" vj < 1 2 g z(r)g ; (x)? " z(x; v) dx R m? " z(x; v) dx R m e? 1 2 jvj2 dv (2) d2 + R z ; jr z j kk P[ jv " j 1 2 g z(r) ] : (2) R m (x)? " z(x; v) dx + R m nm r z and for any x 62 M r z, v 2 D z hence M r z (x)? " z(x; v) dx (x)? " z(x; v) dx ; jh(x)? (z + " v)j jh(x)? zj? j" vj 1 2 g z(r) ;? " z(x; v) exp? 1 8" 2 g2 z(r) ;
5 and j (x)? " z(x; v) dxj R m nmz r kk exp? 1 8" 2 g2 z(r) ; (3) for any v 2 D z. As a consequence, we concentrate below on studying the asymptotic behaviour of z (v) "?d (x)? (2) M " z(x; v) dx ; d2 z r as " # 0, for a given v 2 R d. For the sake of brevity, we will limit ourselves to recover the expression given in the Theorem 2.8 above for a regular value z 2 R d, in the case m d, following the method used in Ellis and Rosen [3], [] and Hwang [7]. Proposition 3. Let Assumption C hold, and assume m d. Let z 2 R d be a regular value, and let z denote the canonical measure on the compact (m? d){ dimensional submanifold. Then, for any xed v 2 R d, and any r > 0 small enough as " # 0. z (v)?! (x) J h (x) z(dx) ; Proof. We prove the result under the additional assumption that the mapping h has a continuous second order derivative. For any x 2, let N x denote the normal space to the submanifold at point x. This is a d{dimensional ane subspace of R m, isomorphic to R d. Dene also N r x f 2 N x : jj < rg ; N r f(x; ) 2 Mz R m : 2 N r x g : Since is compact, we know that for r > 0 small enough, the canonical mapping (x; ) 2 N r 7! x + 2 M r z ; is a dieomorphism, see Berger and Gostiaux [1, Chapter 2]. Moreover, the Jacobian of this dieomorphism is a continuous mapping G z (x; ), with G z (x; 0) 1, for all x 2. The change of variable formula gives z (v) "?d (2) d2 N r x (x + )? " z(x + ; v) G z (x; ) z;x (d) z (dx) ; where z;x is the canonical measure on the d{dimensional linear space N x. To evaluate the inner integral z (x; v) 1 (x + " ) (2) N d2 x r"? " z(x + " ; v) G z (x; " ) z;x (d) ; we use a rst order Taylor expansion for the mapping 7! H " z (x+" ; v), where x 2 and v 2 R d are xed, i.e. H " z(x + " ; v) H " z(x; v) + " [H " z] 0 (x; v) + " (1? ) [H " z ]00 (x + " ; v) d : Notice that if x 2, i.e. if x satises h(x) z, then for every v 2 R d H " z (x; v) 1 2 "2 jvj 2 ; [H " z] 0 (x; v)?" v h 0 (x) ; [H " z] 00 (x; v)?" v h 00 (x) + [h 0 (x)] h 0 (x) : It follows from these observations that 1 " 2 H" z(x + " ; v)?! 1 2 jv? h0 (x) j 2 ; as " # 0. Therefore, by the Lebesgue dominated convergence theorem z (x; v)?! (x) I z (x; v) ; as " # 0, where the integral I z (x; v) is dened as 1 (2) d2 N x exp? 1 2 jv? h0 (x) j 2 z;x (d) : To evaluate this integral we select an orthonormal basis fu 1 ; ; u d g of N x, and we dene the linear mapping U : 2 R d 7! dx k1 This mapping is orthogonal, i.e. k u k 2 N x ' R d : U U IR d and U U I R(U) ; and the change of variable formula gives where I z (x; v) 1 p det (x) ; (x) [h 0 (x) U] h 0 (x) U : Since h 0 (x) U is a square d d matrix, we obtain det (x) det? h 0 (x) U [h 0 (x) U] : Notice that the rows of h 0 (x), i.e. the columns of [h 0 (x)], are d linearly independent m{dimensional vectors in N x R(U). It follows that, for any ; 2 R d h 0 (x) U U [h 0 (x)] h 0 (x) [h 0 (x)] ; since [h 0 (x)] 2 R(U). Therefore p det (x) Jh (x) ; which nishes the proof. 2 Collecting this result with the estimates (2) and (3), we obtain the following limit, i.e. we recover the expression given in the Theorem 2.8 above.
6 Theorem 3.5 Let Assumption C hold, and assume m d. Let z 2 R d be a regular value, and let z denote the canonical measure on the compact (m? d){dimensional submanifold. Then (x) " z?! J h (x) z(dx) ; J h (x) z(dx) as " # 0. APPLICATION TO HYBRID SYSTEMS With the approach presented in this paper, it is possible to address the following estimation problems, which for the sake of simplicity are stated in the static case, i.e. for nite{dimensional r.v.'s rather than for random sequences. Let X be a r.v. in R m, with absolutely continuous probability distribution, and continuous density p w.r.t. the Lebesgue measure on R m. Assume that we learn that X satises the constraint 0 h(x) ; where x 7! h(x) is a proper continuously dierential mapping from R m to R d. If 0 is a regular value of the mapping x 7! h(x), then the conditional probability distribution P [X 2 dx j h(x) 0] can be computed explicitly, and is supported by M fx 2 R m : h(x) 0g. Assume that we learn that X satises instead the constraint 0 g(x; Y ) ; where Y is some observed deterministic variable in R p, e.g. input, parameter, etc., and where x 7! g(x; y) is a proper and continuously dierential mapping from R m to R d, for any y 2 R p. If we observe Y y, and if 0 is a regular value of the mapping x 7! g y (x) g(x; y), then the conditional probability distribution P [X 2 dx j g y (X) 0] can be computed explicitly, and is supported by M y fx 2 R m : g y (x) 0g. 5 EXTENSIONS In the case of a singular value z 2 R d, some partial results could be obtained in the case m d, using ideas contained in Ellis and Rosen [3]. Only the most singular points will contribute to the conditional probability distribution z. The case of observations in continuous time should also be considered. A possible approach would consist to sample the observations and take the limit as the sampling rate goes to innity, but this has not been addressed yet. Let us nally mention another related problem, arising in parameter estimation for a diusion process with small noise. In the case where the standard identi- ability condition does not hold, i.e. where the corresponding Kullback{Leibler information does not have a unique minimum, the consistency of the Bayesian estimate could be obtained as an application of the asymptotics of Laplace integrals. Notice that the case of a nite number of minima of the contrast function has already been considered in Kutoyants [9]. ACKNOWLEDGEMENT The authors gratefully acknowledge Matthew James for bringing the papers [] and [3] to their attention. 6 REFERENCES [1] M. BERGER and B. GOSTIAUX. Dierential Geometry : Manifolds, Curves and Surfaces. Volume 115 of Graduate Texts in Mathematics, Springer{Verlag, New York, [2] L. BREIMAN. Probability. Addison{Wesley, Reading, MA, [3] R.S. ELLIS and J.S. ROSEN. Asymptotic analysis of Gaussian integrals. I : Isolated minimum points. Transactions of the American Mathematical Society, 273(2):7{81, October [] R.S. ELLIS and J.S. ROSEN. Asymptotic analysis of Gaussian integrals, II : Manifold of minimum points. Communications in Mathematical Physics, 82:153{181, [5] L.C. EVANS and R.F. GARIEPY. Measure Theory and Fine Properties of Functions. CRC Press, Boca Raton, FL, [6] H. FEDERER. Geometric Measure Theory. Volume 153 of Grundlehren der mathematischen Wissenschaften, Springer Verlag, Berlin, [7] HWANG Chii{Ruey. Laplace's method revisited : weak convergence of probability measures. The Annals of Probability, 8(6):1177{1182, [8] H. KORELIOGLU and W.J. RUNGGALDIER. Filtering for nonlinear systems driven by nonwhite noises : an approximation scheme. Stochastics and Stochastics Reports, :65{102, [9] Yu.A. KUTOYANTS. Nonconsistent estimation by diusion type observations. Statistics & Probability Letters, 20(1):1{7, May 199. [10] Sh. STERNBERG. Lectures on Dierential Geometry. Prentice Hall, Englewood Clis, NJ, 196. [11] Y. TAKEUCHI and H. AKASHI. Least{squares state estimation of systems with state{dependent observation noise. Automatica, 21(3):303{313, 1985.
Max-Planck-Institut fur Mathematik in den Naturwissenschaften Leipzig Uniformly distributed measures in Euclidean spaces by Bernd Kirchheim and David Preiss Preprint-Nr.: 37 1998 Uniformly Distributed
More informationAbstract. Jacobi curves are far going generalizations of the spaces of \Jacobi
Principal Invariants of Jacobi Curves Andrei Agrachev 1 and Igor Zelenko 2 1 S.I.S.S.A., Via Beirut 2-4, 34013 Trieste, Italy and Steklov Mathematical Institute, ul. Gubkina 8, 117966 Moscow, Russia; email:
More informationLecture 2: Review of Prerequisites. Table of contents
Math 348 Fall 217 Lecture 2: Review of Prerequisites Disclaimer. As we have a textbook, this lecture note is for guidance and supplement only. It should not be relied on when preparing for exams. In this
More informationConverse Lyapunov Functions for Inclusions 2 Basic denitions Given a set A, A stands for the closure of A, A stands for the interior set of A, coa sta
A smooth Lyapunov function from a class-kl estimate involving two positive semidenite functions Andrew R. Teel y ECE Dept. University of California Santa Barbara, CA 93106 teel@ece.ucsb.edu Laurent Praly
More informationN.G.Bean, D.A.Green and P.G.Taylor. University of Adelaide. Adelaide. Abstract. process of an MMPP/M/1 queue is not a MAP unless the queue is a
WHEN IS A MAP POISSON N.G.Bean, D.A.Green and P.G.Taylor Department of Applied Mathematics University of Adelaide Adelaide 55 Abstract In a recent paper, Olivier and Walrand (994) claimed that the departure
More informationLinear Algebra (part 1) : Vector Spaces (by Evan Dummit, 2017, v. 1.07) 1.1 The Formal Denition of a Vector Space
Linear Algebra (part 1) : Vector Spaces (by Evan Dummit, 2017, v. 1.07) Contents 1 Vector Spaces 1 1.1 The Formal Denition of a Vector Space.................................. 1 1.2 Subspaces...................................................
More informationGIOVANNI COMI AND MONICA TORRES
ONE-SIDED APPROXIMATION OF SETS OF FINITE PERIMETER GIOVANNI COMI AND MONICA TORRES Abstract. In this note we present a new proof of a one-sided approximation of sets of finite perimeter introduced in
More informationJ. Huisman. Abstract. let bxc be the fundamental Z=2Z-homology class of X. We show that
On the dual of a real analytic hypersurface J. Huisman Abstract Let f be an immersion of a compact connected smooth real analytic variety X of dimension n into real projective space P n+1 (R). We say that
More information8 Singular Integral Operators and L p -Regularity Theory
8 Singular Integral Operators and L p -Regularity Theory 8. Motivation See hand-written notes! 8.2 Mikhlin Multiplier Theorem Recall that the Fourier transformation F and the inverse Fourier transformation
More informationCHARACTERISTIC POINTS, RECTIFIABILITY AND PERIMETER MEASURE ON STRATIFIED GROUPS
CHARACTERISTIC POINTS, RECTIFIABILITY AND PERIMETER MEASURE ON STRATIFIED GROUPS VALENTINO MAGNANI Abstract. We establish an explicit formula between the perimeter measure of an open set E with C 1 boundary
More informationON TRIVIAL GRADIENT YOUNG MEASURES BAISHENG YAN Abstract. We give a condition on a closed set K of real nm matrices which ensures that any W 1 p -grad
ON TRIVIAL GRAIENT YOUNG MEASURES BAISHENG YAN Abstract. We give a condition on a closed set K of real nm matrices which ensures that any W 1 p -gradient Young measure supported on K must be trivial the
More informationgrowth rates of perturbed time-varying linear systems, [14]. For this setup it is also necessary to study discrete-time systems with a transition map
Remarks on universal nonsingular controls for discrete-time systems Eduardo D. Sontag a and Fabian R. Wirth b;1 a Department of Mathematics, Rutgers University, New Brunswick, NJ 08903, b sontag@hilbert.rutgers.edu
More information2 Garrett: `A Good Spectral Theorem' 1. von Neumann algebras, density theorem The commutant of a subring S of a ring R is S 0 = fr 2 R : rs = sr; 8s 2
1 A Good Spectral Theorem c1996, Paul Garrett, garrett@math.umn.edu version February 12, 1996 1 Measurable Hilbert bundles Measurable Banach bundles Direct integrals of Hilbert spaces Trivializing Hilbert
More informationComputation Of Asymptotic Distribution. For Semiparametric GMM Estimators. Hidehiko Ichimura. Graduate School of Public Policy
Computation Of Asymptotic Distribution For Semiparametric GMM Estimators Hidehiko Ichimura Graduate School of Public Policy and Graduate School of Economics University of Tokyo A Conference in honor of
More informationBROUWER FIXED POINT THEOREM. Contents 1. Introduction 1 2. Preliminaries 1 3. Brouwer fixed point theorem 3 Acknowledgments 8 References 8
BROUWER FIXED POINT THEOREM DANIELE CARATELLI Abstract. This paper aims at proving the Brouwer fixed point theorem for smooth maps. The theorem states that any continuous (smooth in our proof) function
More information1 Brownian Local Time
1 Brownian Local Time We first begin by defining the space and variables for Brownian local time. Let W t be a standard 1-D Wiener process. We know that for the set, {t : W t = } P (µ{t : W t = } = ) =
More information1 Solutions to selected problems
1 Solutions to selected problems 1. Let A B R n. Show that int A int B but in general bd A bd B. Solution. Let x int A. Then there is ɛ > 0 such that B ɛ (x) A B. This shows x int B. If A = [0, 1] and
More informationQUASI-UNIFORMLY POSITIVE OPERATORS IN KREIN SPACE. Denitizable operators in Krein spaces have spectral properties similar to those
QUASI-UNIFORMLY POSITIVE OPERATORS IN KREIN SPACE BRANKO CURGUS and BRANKO NAJMAN Denitizable operators in Krein spaces have spectral properties similar to those of selfadjoint operators in Hilbert spaces.
More informationADJOINTS, ABSOLUTE VALUES AND POLAR DECOMPOSITIONS
J. OPERATOR THEORY 44(2000), 243 254 c Copyright by Theta, 2000 ADJOINTS, ABSOLUTE VALUES AND POLAR DECOMPOSITIONS DOUGLAS BRIDGES, FRED RICHMAN and PETER SCHUSTER Communicated by William B. Arveson Abstract.
More informationThe Erwin Schrodinger International Pasteurgasse 6/7. Institute for Mathematical Physics A-1090 Wien, Austria
ESI The Erwin Schrodinger International Pasteurgasse 6/7 Institute for Mathematical Physics A-1090 Wien, Austria On the Point Spectrum of Dirence Schrodinger Operators Vladimir Buslaev Alexander Fedotov
More informationFENGBO HANG AND PAUL C. YANG
Q CURVATURE ON A CLASS OF 3 ANIFOLDS FENGBO HANG AND PAUL C. YANG Abstract. otivated by the strong maximum principle for Paneitz operator in dimension 5 or higher found in [G] and the calculation of the
More informationECON 7335 INFORMATION, LEARNING AND EXPECTATIONS IN MACRO LECTURE 1: BASICS. 1. Bayes Rule. p(b j A)p(A) p(b)
ECON 7335 INFORMATION, LEARNING AND EXPECTATIONS IN MACRO LECTURE : BASICS KRISTOFFER P. NIMARK. Bayes Rule De nition. Bayes Rule. The probability of event A occurring conditional on the event B having
More informationThe Erwin Schrodinger International Boltzmanngasse 9. Institute for Mathematical Physics A-1090 Wien, Austria
ESI The Erwin Schrodinger International Boltzmanngasse 9 Institute for Mathematical Physics A-19 Wien, Austria The Negative Discrete Spectrum of a Class of Two{Dimentional Schrodinger Operators with Magnetic
More informationContents. 2.1 Vectors in R n. Linear Algebra (part 2) : Vector Spaces (by Evan Dummit, 2017, v. 2.50) 2 Vector Spaces
Linear Algebra (part 2) : Vector Spaces (by Evan Dummit, 2017, v 250) Contents 2 Vector Spaces 1 21 Vectors in R n 1 22 The Formal Denition of a Vector Space 4 23 Subspaces 6 24 Linear Combinations and
More informationON MAPS WHOSE DISTRIBUTIONAL JACOBIAN IS A MEASURE
[version: December 15, 2007] Real Analysis Exchange Summer Symposium 2007, pp. 153-162 Giovanni Alberti, Dipartimento di Matematica, Università di Pisa, largo Pontecorvo 5, 56127 Pisa, Italy. Email address:
More informationContractive metrics for scalar conservation laws
Contractive metrics for scalar conservation laws François Bolley 1, Yann Brenier 2, Grégoire Loeper 34 Abstract We consider nondecreasing entropy solutions to 1-d scalar conservation laws and show that
More informationStochastic Nonlinear Stabilization Part II: Inverse Optimality Hua Deng and Miroslav Krstic Department of Mechanical Engineering h
Stochastic Nonlinear Stabilization Part II: Inverse Optimality Hua Deng and Miroslav Krstic Department of Mechanical Engineering denghua@eng.umd.edu http://www.engr.umd.edu/~denghua/ University of Maryland
More informationBayesian estimation of Hidden Markov Models
Bayesian estimation of Hidden Markov Models Laurent Mevel, Lorenzo Finesso IRISA/INRIA Campus de Beaulieu, 35042 Rennes Cedex, France Institute of Systems Science and Biomedical Engineering LADSEB-CNR
More informationMathematical Institute, University of Utrecht. The problem of estimating the mean of an observed Gaussian innite-dimensional vector
On Minimax Filtering over Ellipsoids Eduard N. Belitser and Boris Y. Levit Mathematical Institute, University of Utrecht Budapestlaan 6, 3584 CD Utrecht, The Netherlands The problem of estimating the mean
More informationSOME MEASURABILITY AND CONTINUITY PROPERTIES OF ARBITRARY REAL FUNCTIONS
LE MATEMATICHE Vol. LVII (2002) Fasc. I, pp. 6382 SOME MEASURABILITY AND CONTINUITY PROPERTIES OF ARBITRARY REAL FUNCTIONS VITTORINO PATA - ALFONSO VILLANI Given an arbitrary real function f, the set D
More informationMarkov processes Course note 2. Martingale problems, recurrence properties of discrete time chains.
Institute for Applied Mathematics WS17/18 Massimiliano Gubinelli Markov processes Course note 2. Martingale problems, recurrence properties of discrete time chains. [version 1, 2017.11.1] We introduce
More informationSHARP BOUNDARY TRACE INEQUALITIES. 1. Introduction
SHARP BOUNDARY TRACE INEQUALITIES GILES AUCHMUTY Abstract. This paper describes sharp inequalities for the trace of Sobolev functions on the boundary of a bounded region R N. The inequalities bound (semi-)norms
More information2 JOSE BURILLO It was proved by Thurston [2, Ch.8], using geometric methods, and by Gersten [3], using combinatorial methods, that the integral 3-dime
DIMACS Series in Discrete Mathematics and Theoretical Computer Science Volume 00, 1997 Lower Bounds of Isoperimetric Functions for Nilpotent Groups Jose Burillo Abstract. In this paper we prove that Heisenberg
More informationSupplement A: Construction and properties of a reected diusion
Supplement A: Construction and properties of a reected diusion Jakub Chorowski 1 Construction Assumption 1. For given constants < d < D let the pair (σ, b) Θ, where Θ : Θ(d, D) {(σ, b) C 1 (, 1]) C 1 (,
More informationSemi-strongly asymptotically non-expansive mappings and their applications on xed point theory
Hacettepe Journal of Mathematics and Statistics Volume 46 (4) (2017), 613 620 Semi-strongly asymptotically non-expansive mappings and their applications on xed point theory Chris Lennard and Veysel Nezir
More informationLECTURE VI: SELF-ADJOINT AND UNITARY OPERATORS MAT FALL 2006 PRINCETON UNIVERSITY
LECTURE VI: SELF-ADJOINT AND UNITARY OPERATORS MAT 204 - FALL 2006 PRINCETON UNIVERSITY ALFONSO SORRENTINO 1 Adjoint of a linear operator Note: In these notes, V will denote a n-dimensional euclidean vector
More informationGeometric Inference for Probability distributions
Geometric Inference for Probability distributions F. Chazal 1 D. Cohen-Steiner 2 Q. Mérigot 2 1 Geometrica, INRIA Saclay, 2 Geometrica, INRIA Sophia-Antipolis 2009 July 8 F. Chazal, D. Cohen-Steiner, Q.
More informationMax-Planck-Institut fur Mathematik in den Naturwissenschaften Leipzig A nonlocal anisotropic model for phase transitions Part II: Asymptotic behaviour of rescaled energies by Giovanni Alberti and Giovanni
More informationThe uniformly accelerated motion in General Relativity from a geometric point of view. 1. Introduction. Daniel de la Fuente
XI Encuentro Andaluz de Geometría IMUS (Universidad de Sevilla), 15 de mayo de 2015, págs. 2934 The uniformly accelerated motion in General Relativity from a geometric point of view Daniel de la Fuente
More informationR. Schaback. numerical method is proposed which rst minimizes each f j separately. and then applies a penalty strategy to gradually force the
A Multi{Parameter Method for Nonlinear Least{Squares Approximation R Schaback Abstract P For discrete nonlinear least-squares approximation problems f 2 (x)! min for m smooth functions f : IR n! IR a m
More informationTechinical Proofs for Nonlinear Learning using Local Coordinate Coding
Techinical Proofs for Nonlinear Learning using Local Coordinate Coding 1 Notations and Main Results Denition 1.1 (Lipschitz Smoothness) A function f(x) on R d is (α, β, p)-lipschitz smooth with respect
More informationStochastic Processes
Introduction and Techniques Lecture 4 in Financial Mathematics UiO-STK4510 Autumn 2015 Teacher: S. Ortiz-Latorre Stochastic Processes 1 Stochastic Processes De nition 1 Let (E; E) be a measurable space
More informationPROOF OF TWO MATRIX THEOREMS VIA TRIANGULAR FACTORIZATIONS ROY MATHIAS
PROOF OF TWO MATRIX THEOREMS VIA TRIANGULAR FACTORIZATIONS ROY MATHIAS Abstract. We present elementary proofs of the Cauchy-Binet Theorem on determinants and of the fact that the eigenvalues of a matrix
More information5.3. Compactness results for Dirichlet spaces Compactness results for manifolds 39 References 39 sec:intro 1. Introduction We mean by a spectr
CONVERGENCE OF SPECTRAL STRUCTURES: A FUNCTIONAL ANALYTIC THEORY AND ITS APPLICATIONS TO SPECTRAL GEOMETRY KAZUHIRO KUWAE AND TAKASHI SHIOYA Abstract. The purpose of this paper is to present a functional
More informationA theorem on summable families in normed groups. Dedicated to the professors of mathematics. L. Berg, W. Engel, G. Pazderski, and H.- W. Stolle.
Rostock. Math. Kolloq. 49, 51{56 (1995) Subject Classication (AMS) 46B15, 54A20, 54E15 Harry Poppe A theorem on summable families in normed groups Dedicated to the professors of mathematics L. Berg, W.
More informationTHE IMPLICIT FUNCTION THEOREM FOR CONTINUOUS FUNCTIONS. Carlos Biasi Carlos Gutierrez Edivaldo L. dos Santos. 1. Introduction
Topological Methods in Nonlinear Analysis Journal of the Juliusz Schauder Center Volume 32, 2008, 177 185 THE IMPLICIT FUNCTION THEOREM FOR CONTINUOUS FUNCTIONS Carlos Biasi Carlos Gutierrez Edivaldo L.
More informationis holomorphic. In other words, a holomorphic function is a collection of compatible holomorphic functions on all charts.
RIEMANN SURFACES 2. Week 2. Basic definitions 2.1. Smooth manifolds. Complex manifolds. Let X be a topological space. A (real) chart of X is a pair (U, f : U R n ) where U is an open subset of X and f
More informationINRIA Rocquencourt, Le Chesnay Cedex (France) y Dept. of Mathematics, North Carolina State University, Raleigh NC USA
Nonlinear Observer Design using Implicit System Descriptions D. von Wissel, R. Nikoukhah, S. L. Campbell y and F. Delebecque INRIA Rocquencourt, 78 Le Chesnay Cedex (France) y Dept. of Mathematics, North
More information4 Derivations of the Discrete-Time Kalman Filter
Technion Israel Institute of Technology, Department of Electrical Engineering Estimation and Identification in Dynamical Systems (048825) Lecture Notes, Fall 2009, Prof N Shimkin 4 Derivations of the Discrete-Time
More information3 Stability and Lyapunov Functions
CDS140a Nonlinear Systems: Local Theory 02/01/2011 3 Stability and Lyapunov Functions 3.1 Lyapunov Stability Denition: An equilibrium point x 0 of (1) is stable if for all ɛ > 0, there exists a δ > 0 such
More informationPointwise convergence rates and central limit theorems for kernel density estimators in linear processes
Pointwise convergence rates and central limit theorems for kernel density estimators in linear processes Anton Schick Binghamton University Wolfgang Wefelmeyer Universität zu Köln Abstract Convergence
More information2 BAISHENG YAN When L =,it is easily seen that the set K = coincides with the set of conformal matrices, that is, K = = fr j 0 R 2 SO(n)g: Weakly L-qu
RELAXATION AND ATTAINMENT RESULTS FOR AN INTEGRAL FUNCTIONAL WITH UNBOUNDED ENERGY-WELL BAISHENG YAN Abstract. Consider functional I(u) = R jjdujn ; L det Duj dx whose energy-well consists of matrices
More informationMicro-support of sheaves
Micro-support of sheaves Vincent Humilière 17/01/14 The microlocal theory of sheaves and in particular the denition of the micro-support is due to Kashiwara and Schapira (the main reference is their book
More informationElementary 2-Group Character Codes. Abstract. In this correspondence we describe a class of codes over GF (q),
Elementary 2-Group Character Codes Cunsheng Ding 1, David Kohel 2, and San Ling Abstract In this correspondence we describe a class of codes over GF (q), where q is a power of an odd prime. These codes
More informationSYMPLECTIC LEFSCHETZ FIBRATIONS ALEXANDER CAVIEDES CASTRO
SYMPLECTIC LEFSCHETZ FIBRATIONS ALEXANDER CAVIEDES CASTRO. Introduction A Lefschetz pencil is a construction that comes from algebraic geometry, but it is closely related with symplectic geometry. Indeed,
More informationSARD S THEOREM ALEX WRIGHT
SARD S THEOREM ALEX WRIGHT Abstract. A proof of Sard s Theorem is presented, and applications to the Whitney Embedding and Immersion Theorems, the existence of Morse functions, and the General Position
More informationWerner Romisch. Humboldt University Berlin. Abstract. Perturbations of convex chance constrained stochastic programs are considered the underlying
Stability of solutions to chance constrained stochastic programs Rene Henrion Weierstrass Institute for Applied Analysis and Stochastics D-7 Berlin, Germany and Werner Romisch Humboldt University Berlin
More informationUniversity of Toronto
A Limit Result for the Prior Predictive by Michael Evans Department of Statistics University of Toronto and Gun Ho Jang Department of Statistics University of Toronto Technical Report No. 1004 April 15,
More informationStabilization as a CW approximation
Journal of Pure and Applied Algebra 140 (1999) 23 32 Stabilization as a CW approximation A.D. Elmendorf Department of Mathematics, Purdue University Calumet, Hammond, IN 46323, USA Communicated by E.M.
More informationOn It^o's formula for multidimensional Brownian motion
On It^o's formula for multidimensional Brownian motion by Hans Follmer and Philip Protter Institut fur Mathemati Humboldt-Universitat D-99 Berlin Mathematics and Statistics Department Purdue University
More informationConvergence of Square Root Ensemble Kalman Filters in the Large Ensemble Limit
Convergence of Square Root Ensemble Kalman Filters in the Large Ensemble Limit Evan Kwiatkowski, Jan Mandel University of Colorado Denver December 11, 2014 OUTLINE 2 Data Assimilation Bayesian Estimation
More informationDefinition 5.1. A vector field v on a manifold M is map M T M such that for all x M, v(x) T x M.
5 Vector fields Last updated: March 12, 2012. 5.1 Definition and general properties We first need to define what a vector field is. Definition 5.1. A vector field v on a manifold M is map M T M such that
More informationSemi-Linear Diusive Representations for Non-Linear Fractional Dierential Systems Jacques AUDOUNET 1, Denis MATIGNON 2?, and Gerard MONTSENY 3 1 MIP /
Semi-Linear Diusive Representations for Non-Linear Fractional Dierential Systems Jacques AUDOUNET 1, Denis MATIGNON?, and Gerard MONTSENY 3 1 MIP / CNRS, Universite Paul Sabatier. 118, route de Narbonne.
More informationSMSTC (2017/18) Geometry and Topology 2.
SMSTC (2017/18) Geometry and Topology 2 Lecture 1: Differentiable Functions and Manifolds in R n Lecturer: Diletta Martinelli (Notes by Bernd Schroers) a wwwsmstcacuk 11 General remarks In this lecture
More informationAbstract. Previous characterizations of iss-stability are shown to generalize without change to the
On Characterizations of Input-to-State Stability with Respect to Compact Sets Eduardo D. Sontag and Yuan Wang Department of Mathematics, Rutgers University, New Brunswick, NJ 08903, USA Department of Mathematics,
More informationLie Groups for 2D and 3D Transformations
Lie Groups for 2D and 3D Transformations Ethan Eade Updated May 20, 2017 * 1 Introduction This document derives useful formulae for working with the Lie groups that represent transformations in 2D and
More informationMath 350 Fall 2011 Notes about inner product spaces. In this notes we state and prove some important properties of inner product spaces.
Math 350 Fall 2011 Notes about inner product spaces In this notes we state and prove some important properties of inner product spaces. First, recall the dot product on R n : if x, y R n, say x = (x 1,...,
More informationMath 413/513 Chapter 6 (from Friedberg, Insel, & Spence)
Math 413/513 Chapter 6 (from Friedberg, Insel, & Spence) David Glickenstein December 7, 2015 1 Inner product spaces In this chapter, we will only consider the elds R and C. De nition 1 Let V be a vector
More information5 Banach Algebras. 5.1 Invertibility and the Spectrum. Robert Oeckl FA NOTES 5 19/05/2010 1
Robert Oeckl FA NOTES 5 19/05/2010 1 5 Banach Algebras 5.1 Invertibility and the Spectrum Suppose X is a Banach space. Then we are often interested in (continuous) operators on this space, i.e, elements
More informationMathematics Department Stanford University Math 61CM/DM Inner products
Mathematics Department Stanford University Math 61CM/DM Inner products Recall the definition of an inner product space; see Appendix A.8 of the textbook. Definition 1 An inner product space V is a vector
More informationA quasi-separation theorem for LQG optimal control with IQ constraints Andrew E.B. Lim y & John B. Moore z February 1997 Abstract We consider the dete
A quasi-separation theorem for LQG optimal control with IQ constraints Andrew E.B. Lim y & John B. Moore z February 1997 Abstract We consider the deterministic, the full observation and the partial observation
More informationREMARKS ON THE TIME-OPTIMAL CONTROL OF A CLASS OF HAMILTONIAN SYSTEMS. Eduardo D. Sontag. SYCON - Rutgers Center for Systems and Control
REMARKS ON THE TIME-OPTIMAL CONTROL OF A CLASS OF HAMILTONIAN SYSTEMS Eduardo D. Sontag SYCON - Rutgers Center for Systems and Control Department of Mathematics, Rutgers University, New Brunswick, NJ 08903
More informationIndependence of some multiple Poisson stochastic integrals with variable-sign kernels
Independence of some multiple Poisson stochastic integrals with variable-sign kernels Nicolas Privault Division of Mathematical Sciences School of Physical and Mathematical Sciences Nanyang Technological
More informationMethods for a Class of Convex. Functions. Stephen M. Robinson WP April 1996
Working Paper Linear Convergence of Epsilon-Subgradient Descent Methods for a Class of Convex Functions Stephen M. Robinson WP-96-041 April 1996 IIASA International Institute for Applied Systems Analysis
More informationLinear Algebra. Linear Algebra. Chih-Wei Yi. Dept. of Computer Science National Chiao Tung University. November 12, 2008
Linear Algebra Chih-Wei Yi Dept. of Computer Science National Chiao Tung University November, 008 Section De nition and Examples Section De nition and Examples Section De nition and Examples De nition
More informationA slow transient diusion in a drifted stable potential
A slow transient diusion in a drifted stable potential Arvind Singh Université Paris VI Abstract We consider a diusion process X in a random potential V of the form V x = S x δx, where δ is a positive
More informationthe set of critical points of is discrete and so is the set of critical values of. Suppose that a 1 < b 1 < c < a 2 < b 2 and let c be the only critic
1. The result DISCS IN STEIN MANIFOLDS Josip Globevnik Let be the open unit disc in C. In the paper we prove the following THEOREM Let M be a Stein manifold, dimm 2. Given a point p 2 M and a vector X
More informationLecture 2: From Linear Regression to Kalman Filter and Beyond
Lecture 2: From Linear Regression to Kalman Filter and Beyond January 18, 2017 Contents 1 Batch and Recursive Estimation 2 Towards Bayesian Filtering 3 Kalman Filter and Bayesian Filtering and Smoothing
More informationA Representation of Excessive Functions as Expected Suprema
A Representation of Excessive Functions as Expected Suprema Hans Föllmer & Thomas Knispel Humboldt-Universität zu Berlin Institut für Mathematik Unter den Linden 6 10099 Berlin, Germany E-mail: foellmer@math.hu-berlin.de,
More informationON THE DIAMETER OF THE ATTRACTOR OF AN IFS Serge Dubuc Raouf Hamzaoui Abstract We investigate methods for the evaluation of the diameter of the attrac
ON THE DIAMETER OF THE ATTRACTOR OF AN IFS Serge Dubuc Raouf Hamzaoui Abstract We investigate methods for the evaluation of the diameter of the attractor of an IFS. We propose an upper bound for the diameter
More information(1.) For any subset P S we denote by L(P ) the abelian group of integral relations between elements of P, i.e. L(P ) := ker Z P! span Z P S S : For ea
Torsion of dierentials on toric varieties Klaus Altmann Institut fur reine Mathematik, Humboldt-Universitat zu Berlin Ziegelstr. 13a, D-10099 Berlin, Germany. E-mail: altmann@mathematik.hu-berlin.de Abstract
More informationSobolev spaces. May 18
Sobolev spaces May 18 2015 1 Weak derivatives The purpose of these notes is to give a very basic introduction to Sobolev spaces. More extensive treatments can e.g. be found in the classical references
More informationFinite-dimensional spaces. C n is the space of n-tuples x = (x 1,..., x n ) of complex numbers. It is a Hilbert space with the inner product
Chapter 4 Hilbert Spaces 4.1 Inner Product Spaces Inner Product Space. A complex vector space E is called an inner product space (or a pre-hilbert space, or a unitary space) if there is a mapping (, )
More informationConvergence of the Ensemble Kalman Filter in Hilbert Space
Convergence of the Ensemble Kalman Filter in Hilbert Space Jan Mandel Center for Computational Mathematics Department of Mathematical and Statistical Sciences University of Colorado Denver Parts based
More informationp (z) = p z 1 pz : The pseudo-hyperbolic distance ½(z; w) between z and w in D is de ned by ½(z; w) = j z (w)j =
EXTREME POINTS OF THE CLOSED CONVEX HULL OF COMPOSITION OPERATORS TAKUYA HOSOKAW A Abstract. W e study the extreme points of the closed convex hull of the set of all composition operators on the space
More informationAlgebraic Information Geometry for Learning Machines with Singularities
Algebraic Information Geometry for Learning Machines with Singularities Sumio Watanabe Precision and Intelligence Laboratory Tokyo Institute of Technology 4259 Nagatsuta, Midori-ku, Yokohama, 226-8503
More informationOn Coarse Geometry and Coarse Embeddability
On Coarse Geometry and Coarse Embeddability Ilmari Kangasniemi August 10, 2016 Master's Thesis University of Helsinki Faculty of Science Department of Mathematics and Statistics Supervised by Erik Elfving
More informationg(.) 1/ N 1/ N Decision Decision Device u u u u CP
Distributed Weak Signal Detection and Asymptotic Relative Eciency in Dependent Noise Hakan Delic Signal and Image Processing Laboratory (BUSI) Department of Electrical and Electronics Engineering Bogazici
More information2 Section 2 However, in order to apply the above idea, we will need to allow non standard intervals ('; ) in the proof. More precisely, ' and may gene
Introduction 1 A dierential intermediate value theorem by Joris van der Hoeven D pt. de Math matiques (B t. 425) Universit Paris-Sud 91405 Orsay Cedex France June 2000 Abstract Let T be the eld of grid-based
More informationGarrett: `Bernstein's analytic continuation of complex powers' 2 Let f be a polynomial in x 1 ; : : : ; x n with real coecients. For complex s, let f
1 Bernstein's analytic continuation of complex powers c1995, Paul Garrett, garrettmath.umn.edu version January 27, 1998 Analytic continuation of distributions Statement of the theorems on analytic continuation
More informationz x = f x (x, y, a, b), z y = f y (x, y, a, b). F(x, y, z, z x, z y ) = 0. This is a PDE for the unknown function of two independent variables.
Chapter 2 First order PDE 2.1 How and Why First order PDE appear? 2.1.1 Physical origins Conservation laws form one of the two fundamental parts of any mathematical model of Continuum Mechanics. These
More informationsemi-group and P x the law of starting at x 2 E. We consider a branching mechanism : R +! R + of the form (u) = au + bu 2 + n(dr)[e ru 1 + ru]; (;1) w
M.S.R.I. October 13-17 1997. Superprocesses and nonlinear partial dierential equations We would like to describe some links between superprocesses and some semi-linear partial dierential equations. Let
More informationSYMPLECTIC GEOMETRY: LECTURE 5
SYMPLECTIC GEOMETRY: LECTURE 5 LIAT KESSLER Let (M, ω) be a connected compact symplectic manifold, T a torus, T M M a Hamiltonian action of T on M, and Φ: M t the assoaciated moment map. Theorem 0.1 (The
More informationMetric Thickenings of Euclidean Submanifolds
Metric Thickenings of Euclidean Submanifolds Advisor: Dr. Henry Adams Committe: Dr. Chris Peterson, Dr. Daniel Cooley Joshua Mirth Masters Thesis Defense October 3, 2017 Introduction Motivation Can we
More informationf(z)dz = 0. P dx + Qdy = D u dx v dy + i u dy + v dx. dxdy + i x = v
MA525 ON CAUCHY'S THEOREM AND GREEN'S THEOREM DAVID DRASIN (EDITED BY JOSIAH YODER) 1. Introduction No doubt the most important result in this course is Cauchy's theorem. Every critical theorem in the
More informationMath 321: Linear Algebra
Math 32: Linear Algebra T. Kapitula Department of Mathematics and Statistics University of New Mexico September 8, 24 Textbook: Linear Algebra,by J. Hefferon E-mail: kapitula@math.unm.edu Prof. Kapitula,
More information1 Introduction It will be convenient to use the inx operators a b and a b to stand for maximum (least upper bound) and minimum (greatest lower bound)
Cycle times and xed points of min-max functions Jeremy Gunawardena, Department of Computer Science, Stanford University, Stanford, CA 94305, USA. jeremy@cs.stanford.edu October 11, 1993 to appear in the
More informationWeek 2 Spring Lecture 3. The Canonical normal means estimation problem (cont.).! (X) = X+ 1 X X, + (X) = X+ 1
Week 2 Spring 2009 Lecture 3. The Canonical normal means estimation problem (cont.). Shrink toward a common mean. Theorem. Let X N ; 2 I n. Let 0 < C 2 (n 3) (hence n 4). De ne!! (X) = X+ 1 Then C 2 X
More informationCOMPACT DIFFERENCE OF WEIGHTED COMPOSITION OPERATORS ON N p -SPACES IN THE BALL
COMPACT DIFFERENCE OF WEIGHTED COMPOSITION OPERATORS ON N p -SPACES IN THE BALL HU BINGYANG and LE HAI KHOI Communicated by Mihai Putinar We obtain necessary and sucient conditions for the compactness
More informationKKT Conditions for Rank-Deficient Nonlinear Least-Square Problems with Rank-Deficient Nonlinear Constraints1
JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS: Vol. 100, No. 1. pp. 145-160, JANUARY 1999 KKT Conditions for Rank-Deficient Nonlinear Least-Square Problems with Rank-Deficient Nonlinear Constraints1
More information