On Dynkin's Markov property of random fields associated with symmetric processes

Let p(t, x, y) be a symmetric transition density with respect to a σ-finite measure m on (E, $\text{E}$), g(x,y)=∫p(t,x,y)dt, and $\text{M}\text{={σ-finite measures μ?0:∫g(x,y)μ(}\text{d}\text{x)μ(}\text{d}\text{y)<∞}}$. There exists a Gaussian random field $\text{Φ={?}{}_{\mathrm{\mu }}\text{:μ?}\text{M}\text{}}$ with mean 0 and covariance $\text{E}\text{?}{}_{\mathrm{\mu }}\text{?}{}_{\mathrm{\nu }}\text{=∫g(x,y)μ(}\text{d}\text{x)ν(}\text{d}\text{y)}$. Letting $\text{F}\text{(B)=σ{?}{}_{\mathrm{\mu }}\text{:μ(B}{}^{\mathrm{c}}\text{)=0}}$ we consider necessary and sufficient conditions for the Markov property (MP) on sets B, C: $\text{F}$(B), $\text{F}$(C) c.i. given $\text{F}$(B ∩ C). Of crucial importance is the following, proved by Dynkin: $\text{E}\text{{?}{}_{\mathrm{\mu }}\text{∣}\text{F}\text{(B)}=?}{}_{\mathrm{\mu B}}$, where μ_B is the hitting distribution of the process corresponding to p, m with initial law μ. Another important fact is that ?_μ=?_ν iff μ, ν have the same potential. Putting these together with an additional transience assumption, we present a potential theoretic proof of the following necessary and sufficient condition for (MP) on sets B, C: For every x?E, T_B∩C=T_B+T_C∮ θ_TB=T_C+T_B∮θ_TC a.s. P^x where, for D ? $\text{E}$, T_D is the hitting time of D for the process associated with p, m. This implies a necessary condition proved by Dynkin in a recent preprint for the case where B∪C=E and B, C are finely closed.     

Partitions of large multipartites with congruence conditions. II

For 1 ≦ l ≦ j, let $\text{a}$_l = ?_h=1^q(l){a_lh + Mv: v = 0, 1, 2,…}, where j, M, q(l) and the a_lh are positive integers such that j > 1, a_l1 < … < a_lq(2) ≦ M, and let $\text{a}$_l^′ = $\text{a}$_l ∪ {0}. Let p(n : $\text{B}$) be the number of partitions of n = (n₁,…,n_j) where, for 1 ≦ l ≦ j, the lth component of each part belongs to $\text{B}$_l and let p¹(n : $\text{B}$) be the number of partitions of n into different parts where again the lth component of each part belongs to $\text{B}$_l. Asymptotic formulas are obtained for p(n : $\text{a}$), p¹(n : $\text{a}$) where all but one n_l tend to infinity much more rapidly than that n_l, and asymptotic formulas are also obtained for p(n : $\text{a}$′), p¹(n ; $\text{a}$′), where one n_l tends to infinity much more rapidly than every other n_l. These formulas contrast with those of a recent paper (Robertson and Spencer, Trans. Amer. Math. Soc., to appear) in which all the n_l tend to infinity at approximately the same rate.     

A new fuzzy compactness defined by fuzzy nets

Let (Ω, $\text{A}$, μ) be a probability space and let $\text{B}$ be a subsigma algebra of $\text{A}$. Let A= L_∞Ω, $\text{A}$, μ , let A= L_∞Ω, $\text{B}$, μ, and let f?A. It is shown that best L_∞-approximations of f by elements of B comprise an interval in B; that is, there exists $\text{f}\text{,}\text{f}\text{?B}$ such that a function g?B is a best L_∞-approximation to f if and only if $\text{f}\text{? g ?}\text{f}$ a.e. on Ω. The difference, $\text{f}\text{? f}$, of $\text{f}$ and f is completely characterized in terms of special sets that have been developed in [2]. Then it is established that the best best L_∞-approximation, f_{$\text{B}$,∞}, to f by elements of B is the average of $\text{f}$ and $\text{f}$, where the function f_{$\text{B}$,∞} is defined by f_{$\text{B}$,∞}(ω) lim_{p → ∞}f_$\text{B}$,P(ξ) and f_$\text{B}$,P denotes the best L_p-approximation to f elements of L_p(Ω, $\text{B}$, μ).     

The inverse balayage problem for Markov chains,part II

This paper continues the study of the inverse balayage problem for Markov chains. Let X be a Markov chain with state space $\text{A ? B}{}_2$, let v be a probability measure on B₂ and let M(v) consist of probability measures μ on A whose X-balayage onto B₂ is v. The faces of the compact, convex set M(v) are characterized. For fixed μ?M(v) the set M(μ,v) of the measures ? of the form $\text{?(·) = P}{}^{\mathrm{\mu }}\text{{X(S) ? ·}}$, where S is a randomized stopping time, is analyzed in detail. In particular, its extreme points and edge are explicitly identified. A naturally defined reversed chain X, for which v is an inverse balayage of μ, is introduced and the relation between X and X^ is studied. The question of which $\text{? ? M(μ, v)}$ admit a natural stopping time $\text{S}{}_{\mathrm{?}}$ of X (not involving an independent randomization) such that $\text{?(·) = P}{}^{\mathrm{\mu }}\text{{X(S}{}_{\mathrm{?}}\text{) ? ·}}$, is shown to have rather different answers in discrete and continuous time. Illustrative examples are presented.     

A note on the conditional moments of a multivariate normal distribution confined to a convex set

Let Y be an N(μ, Σ) random variable on R^m, 1 ≤ m ≤ ∞, where Σ is positive definite. Let C be a nonempty convex set in R^m with closure $\text{C}$. Let (·,-·) be the Eculidean inner product on R^m, and let μ_c be the conditional expected value of Y given Y ∈ C. For v ∈ R^m and s ≥ 0, let β_s(v) be the expected value of |(v, Y) ? (v, μ)|^s and let γ_s(v) be the conditional expected value of |(v, Y) ? (v, μ_c)|^s given Y ∈ C. For s ≥ 1, γ_s(v) < β_s(v) if and only if $\text{C}\text{+ Σ v ≠}\text{C}$, and γ_s(v) < β_s(v) for all v ≠ 0 if and only if $\text{C}\text{+ v ≠}\text{C}$ for any v ∈ R^m such that v ≠ 0.     

Classifying Hilbert bundles. II

A Hilbert bundle (p, B, X) is a type of fibre space p: B → X such that each fibre p^?1(x) is a Hilbert space. However, p^?1(x) may vary in dimension as x varies in X, even when X is connected. We give two “homotopy” type classification theorems for Hilbert bundles having primarily finite dimensional fibres. An (m, n)-bundle over the pair (X, A) is a Hilbert bundle over (p, B, X) such that the dimension of p^?1(x) is m for x in A and n otherwise. As a special case, we show that if X is a compact metric space, C₊X the upper cone of the suspension SX, then the isomorphism classes of (m, n)-bundles over (SX, C₊X) are in one-to-one correspondence with the members of [X, V_m($\text{C}$ⁿ)] where V_m($\text{C}$ⁿ) is the Stiefel manifold. The results are all applicable to the classification of separable, continuous trace C¹-algebras, with specific results given to illustrate.     

On the absolute continuity of Schrödinger operators with spherically symmetric,long-range potentials,I

Let H = ?Δ + V, where the potential V is spherically symmetric and can be decomposed as a sum of a short-range and a long-range term, V(r) = V_S(r) + V_L(r). Assume that for some r₀, V_L(r) ?C^2k(r₀, ∞) and that there exist μ > 0, δ > 0, such that $\text{(}\text{d}\text{dr}\text{)}{}^{\mathrm{j}}\text{V}{}_{\mathrm{<mtext>L</mtext>}}\text{(r) = O(r}{}^{\mathrm{?\mu ?j\delta }}\text{)}\text{as}\text{r → ∞, 1 ? j ? 2k}$. Assume further that min(2kμ, (2k ? 1)δ + μ) > 1. Under this weak decay condition on V_L(r) it is shown in this paper that the positive spectrum of H is absolutely continuous and that the absolutely continuous part of H is unitarily equivalent to ?Δ, provided that the singularity of V at 0 is properly restricted. In particular, some oscillation of V_L(r) at infinity is allowed.     

Unitarization of symplectics and stability for causal differential equations in Hilbert space

Let Sp($\text{H}$) be the symplectic group for a complex Hibert space $\text{H}$. Its Lie algebra sp($\text{H}$) contains an open invariant convex cone C₀; each element of C₀ commutes with a unique sympletic complex structure. The Cayley transform $\text{C}$: X∈ sp($\text{H}$)→(I + X)¹∈ Sp($\text{H}$) is analyzed and compared with the exponential mapping. As an application we consider equations of the form $\text{(}\text{d}\text{dt}\text{) S = A(t)S,}\text{where}\text{t → A(t) ?}\text{C}\text{?}{}_0$ is strongly continuous, and show that if ∝_?∞^∞ ∥A(t)∥ dt < 2 and ∝_{? t8}^∞A(t) dt?C₀, the (scattering) operator

, where S_t′(t) is the solution such that S_t′(t′) = I, is in the range of $\text{B}$ restricted to C₀. It follows that S leaves invariant a unique complex structure; in particular, it is conjugate in Sp($\text{H}$) to a unitary operator.     

K-theory for the leaf space of foliations by Reeb components

The K-theory of the C¹-algebra $\text{C}{}^1\text{(V, F)}$ associated to C^∞-foliations (V, F) of a manifold V in the simplest non-trivial case, i.e., dim V = 2, is studied. Since the case of the Kronecker foliation was settled by Pimsner and Voiculescu (J. Operator Theory4 (1980), 93–118), the remaining problem deals with foliations by Reeb components. The K-theory of $\text{C}{}^1\text{(V, F)}$ for the Reeb foliation of S³ is also computed. In these cases the C¹-algebra $\text{C}{}^1\text{(V, F)}$ is obtained from simpler C¹-algebras by means of pullback diagrams and short exact sequences. The K-groups $\text{K}{}_1\text{(C}{}^1\text{(V, F))}$ are computed using the associated Mayer-Vietoris and six-term exact sequences. The results characterize the C¹-algebra of the Reeb foliation of $\text{T}$² uniquely as an extension of C(S¹) by C(S¹). For the foliations of $\text{T}$² it is found that the K-groups count the number of Reeb components separated by stable compact leaves. A C^∞-foliation of $\text{T}$² such that K₁(C¹($\text{T}$², F)) has infinite rank is also constructed. Finally it is proved, by explicit calculation using (M. Penington, “K-Theory and C¹-Algebras of Lie Groups and Foliations,” D. Phil. thesis, Oxford, 1983), that the natural map $\text{μ: K}{}_1\text{,}{}_{\mathrm{\tau }}\text{(BG) → K}{}_1\text{(C}{}^1\text{(V, F))}$ is an isomorphism for foliations by Reeb components of $\text{T}$² and S³. In particular this proves the Baum-Connes conjecture (P. Baum and A. Connes, Geometric K-theory for Lie groups, preprint, 1982; A. Connes, Proc. Symp. Pure Math.38 (1982), 521–628) when V = $\text{T}$².     

A global genericity theorem for bifurcations in variational problems

Let C be a Banach space, H a Hilbert space, and let F(C,H) be the space of C^∞ functions f: C × H → R having Fredholm second derivative with respect to x at each (c, x) ?C × H for which $\text{D?}{}_{\mathrm{c}}\text{(x) = 0}$; here we write $\text{?}{}_{\mathrm{c}}\text{(x)}$ for $\text{?(c, x)}$. Say ? is of standard type if at all critical points of ?_c it is locally equivalent (as an unfolding) to a quadratic form Q plus an elementary catastrophe on the kernel of Q. It is proved that if f?F (A × B, H) satisfies a certain ‘general position’ condition, and dim B ? 5, then for most a?A the function f_o?F(B,H) is of standard type. Using this it is shown that those f?F(B,H) of standard type form an open dense set in F(B,H) with the Whitney topology. Thus both results are Hilbert-space versions of Thom's theorem for catastrophes in $\text{R}$ⁿ.     

Injectivity and operator spaces

Let B(H) be the bounded operators on a Hilbert space H. A linear subspace R ? B(H) is said to be an operator system if 1 ?R and R is self-adjoint. Consider the category $\text{b}$ of operator systems and completely positive linear maps. R ∈ $\text{C}$ is said to be injective if given A ? B, A, B ∈ $\text{C}$, each map A → R extends to B. Then each injective operator system is isomorphic to a conditionally complete C¹-algebra. Injective von Neumann algebras R are characterized by any one of the following: (1) a relative interpolation property, (2) a finite “projectivity” property, (3) letting M_m = B(C^m), each map R → N ? M_m has approximate factorizations R → M_n → N, (4) letting K be the orthogonal complement of an operator system N ? M_m, each map $\text{M}{}_{\mathrm{m}}\text{K}\text{→ R}$ has approximate factorizations $\text{M}{}_{\mathrm{m}}\text{K}\text{→ M}{}_{\mathrm{n}}\text{→ R}$. Analogous characterizations are found for certain classes of C¹-algebras.     

On convergence of LAD estimates in autoregression with infinite variance

The least absolute deviation estimates L(N), from N data points, of the autoregressive constants a = (a₁, …, a_q)′ for a stationary autoregressive model, are shown to have the property that N^σ(L(N) ? a) converge to zero in probability, for $\text{σ <}\text{1}\text{α}$, where the disturbances are i.i.d., attracted to a stable law of index α, 1 ≤ α < 2, and satisfy some other conditions.     

Integral formulas with distribution kernels for irreducible projections in L2 of a nilmanifold

Let N be a simply connected nilpotent Lie group and Γ a discrete uniform subgroup. The authors consider irreducible representations σ in the spectrum of the quasi-regular representation N × L²(Γ/N) → L²(Γ→) which are induced from normal maximal subordinate subgroups M ? N. The primary projection P_σ and all irreducible projections P ? P_σ are given by convolutions involving right Γ-invariant distributions D on Γ→, $\text{Pf(Γn) = D 1 f(Γn) = <D, n · f>}\text{all}\text{f ? C}{}^{\mathrm{\infty }}\text{(Γ/N)}$, where n · f(ζ) = f(ζ · n). Extending earlier work of Auslander and Brezin, and L. Richardson, the authors give explicit character formulas for the distributions, interpreting them as sums of characters on the torus T^κ = (Γ ∩ M) · [M, M]?M. By examining these structural formulas, they obtain fairly sharp estimates on the order of the distributions: if σ is associated with an orbit $\text{O}$ ? $\text{n}$¹ and if $\text{V}$ ? $\text{n}$¹ is the largest subspace which saturates θ in the sense that f ? $\text{O}$ ? f + $\text{V}$ ? $\text{O}$. As a corollary they obtain Richardson's criterion for a projection to map C⁰(Γ→) into itself. The authors also resolve a conjecture of Brezin, proving a Zero-One law which says, among other things, that if the primary projection P_σ maps C^r(Γ→) into C⁰(Γ→), so do all irreducible projections P ? P_σ. This proof is based on a classical lemma on the extent to which integral points on a polynomial graph in Rⁿ lie in the coset ring of Zⁿ (the finitely additive Boolean algebra generated by cosets of subgroups in Zⁿ). This lemma may be useful in other investigations of nilmanifolds.     

Unbounded operators: Perturbations and commutativity problems

Let H be a complex Hilbert space, and let G_i, i = 1, 2, be closed and orthogonal subspaces of the product space H × H. The subspace G = G₁ ⊕ G₂ is called a (graph) perturbation. We give conditions for invariance of regular operators (R.O.) under graph perturbations: When is the perturbation of a R.O. again a R.O.? If N is a Hilbert space we consider R.O. (i.e., densely defined and closed operators T) in H=$\text{L}$(N) such that G(T)=G(S)⊕V$\text{H}$(M, where G denotes the graph, S is a decomposable operator in H, V a decomposable partial isometry such that the initial space of V(t) is equal to M a.e. t, and finally $\text{H}$(M) is the Hardy space of analytic $\text{L}$₂ vector functions with values in M ? N × N. Such operators T commute with the bilateral shift U; but, unless M = 0, T does not commute with $\text{U}{}^1$. Conversely, this is a canonical model for all R.O. with said commutativity properties. Moreover, the model is unique when T is given, and M = G(w) where w is a partial isometry in N. The detailed structure of the model is analyzed in the special case where dim N = dim M = 1. We relate the problem to a condition of Szeg? by showing that T is a $\text{R.O. iff}\text{∝}{}_0{}^{\mathrm{2\pi }}\text{log}\text{¦ V}{}_2\text{(t)¦ dt = ?∞,}\text{where}\text{V = (V}{}_1\text{, V}{}_2\text{)}$ is the partial isometry in the special case of dimension one. Szeg?'s conditions enters in a different way in the analysis of the case M = N × N, as well as in the spectral analysis of T. Our results provide an answer to a commutativity problem posed by Fuglede. If T is an arbitrary densely defined operator, and A?B(H) is normal, we prove two theorems stating conditions for the implication $\text{A}{}_{\mathrm{\cup }}\text{T ? A}{}_{\mathrm{\cup }}{}^1\text{T}$. These conditions cannot generally be relaxed.     

The visually distinct configurations of k sets

Two sets of sets, C₀ and C₁, are said to be visually equivalent if there is a 1-1 mapping m from C₀ onto C₁ such that for every S, T?C₀, S ∩ T=0 if and only if m(S)∩ m(T)=0 and S?T if and only if m(S)?m(T). We find estimates for V(k), the number of equivalence classes of this relation on sets of k sets, for finite and infinite k. Our main results are that for finite k, $\text{1}\text{2}\text{k}{}^2\text{-k log k <log V (k)<ak}{}^2\text{+βk+log k}$, where α and β are approximately 0.7255 and 2.5323 respectively, and there is a set N of cardinality $\text{1}\text{2}\text{(k}{}^2\text{+k)}$ such that there are V(k) visually distinct sets of k subsets of N.     

Construction of translation planes from t-spread sets

A t-spread set [1] is a set $\text{C}$ of (t + 1) × (t + 1) matrices over GF(q) such that ∥C∥ = q^t+1, 0 ? C, I?C, and det(X ? Y) ≠ 0 if X and Y are distinct elements of $\text{C}$. The amount of computation involved in constructing t-spread sets is considerable, and the following construction technique reduces somewhat this computation. Construction: Let $\text{G}$ be a subgroup of GL(t + 1, q), (the non-singular (t + 1) × (t + 1) matrices over GF(q)), such that ∥G∥|a^t+1, and det (G ? H) ≠ 0 if G and H are distinct elements of $\text{G}$. Let A₁, A₂, …, A_n?GL(t + 1, q) such that det(A_i ? G) ≠ 0 for i = 1, …, n and all G?G, and det(A_i ? A_jG) ≠ 0 for i > j and all G?G. Let $\text{C = \&{0\&} ∪ G ∪ A}{}_1\text{G ∪ … ∪ A}{}_{\mathrm{n}}\text{G}$, and ∥C∥ = q^t+1. Then $\text{C}$ is a t-spread set. A t-spread set can be used to define a left V ? W system over V(t + 1, q) as follows: x + y is the vector sum; let e?V(t + 1, q), then xoy = yM(x) where M(x) is the unique element of $\text{C}$ with x = eM(x). Theorem: Let$\text{C}$be a t-spread set and F the associatedV ? Wsystem; the left nucleus = {y | CM(y) = C}, and the middle nucleus = }y | M(y)C = C}. Theorem: For$\text{C}$constructed as aboveG ? {M(x) | x?N_λ}. This construction technique has been applied to construct a V ? W system of order 25 with ∥N_λ∥ = 6, and ∥N_μ∥ = 4. This system coordinatizes a new projective plane.     

Calculation of the Shannon information

Let ($\text{Ω, β, μ}{}_{\mathrm{X}}$) and $\text{(?, F, μ}{}_{\mathrm{N}}\text{)}$ be probability spaces, with $\text{f: Ω × ? ? ?}\text{a}\text{β × F|F}$ measurable map. Define μ_XY on β × $\text{F}$ by μ_XY(A) = μ_X ? μ_N{(x, y): (x, f(x, y)) ?A}, and let μ_Y = (μ_X ? μ_N)^of^?1. An expression is determined for computing the Shannon information in the measure μ_XY. This expression is used to compute the information for the non-linear additive Gaussian channel, and can be used to solve the channel capacity problem.     

Planar wiener inverses

Let B(Z₁, z₂) be a power series and C(z₁, z₂) be the least mean-square inverse approximation of B among polynomials of a fixed order. This paper discusses conditions under which $\text{B}\text{?}\text{(z}{}_1\text{, z}{}_2\text{)}$, the least mean-square inverse polynomial approximation of C, does not vanish on the unit bicylinder.