共查询到20条相似文献,搜索用时 31 毫秒
1.
Let be a C*-dynamical system where be a semigroup of injective endomorphism and ψ be an (λt) invariant state on the C* subalgebra and is either non-negative integers or real numbers. The central aim of this exposition is to find a useful criteria for the inductive limit state canonically associated with ψ to be pure. We achieve this by exploring the minimal weak forward and backward Markov processes associated with the Markov semigroup on the corner von-Neumann algebra of the support projection of the state ψ to prove that Kolmogorov's property [A. Mohari, Markov shift in non-commutative probability, J. Funct. Anal. 199 (2003) 189–209] of the Markov semigroup is a sufficient condition for the inductive state to be pure. As an application of this criteria we find a sufficient condition for a translation invariant factor state on a one-dimensional quantum spin chain to be pure. This criteria in a sense complements criteria obtained in [O. Bratteli, P.E.T. Jorgensen, A. Kishimoto, R.F. Werner, Pure states on , J. Operator Theory 43 (1) (2000) 97–143; A. Mohari, Markov shift in non-commutative probability, J. Funct. Anal. 199 (2003) 189–209] as we could go beyond lattice symmetric states. 相似文献
2.
Some rigidity results for non-commutative Bernoulli shifts 总被引:3,自引:0,他引:3
We introduce the outer conjugacy invariants , for cocycle actions σ of discrete groups G on type II1 factors N, as the set of real numbers t>0 for which the amplification σt of σ can be perturbed to an action, respectively, to a weakly mixing action. We calculate explicitly and the fundamental group of σ, , in the case G has infinite normal subgroups with the relative property (T) (e.g., when G itself has the property (T) of Kazhdan) and σ is an action of G on the hyperfinite II1 factor by Connes–Størmer Bernoulli shifts of weights {ti}i. Thus, and coincide with the multiplicative subgroup S of generated by the ratios {ti/tj}i,j, while if S={1} (i.e. when all weights are equal), and otherwise. In fact, we calculate all the “1-cohomology picture” of σt,t>0, and classify the actions (σ,G) in terms of their weights {ti}i. In particular, we show that any 1-cocycle for (σ,G) vanishes, modulo scalars, and that two such actions are cocycle conjugate iff they are conjugate. Also, any cocycle action obtained by reducing a Bernoulli action of a group G as above on to the algebra pNp, for p a projection in N, p≠0,1, cannot be perturbed to a genuine action. 相似文献
3.
It is shown that for the inclusion of factors corresponding to an inclusion of ergodic discrete measured equivalence relations , is normal in in the sense of Feldman–Sutherland–Zimmer [J. Feldman, C.E. Sutherland, R.J. Zimmer, Subrelations of ergodic equivalence relations, Ergodic Theory Dynam. Systems 9 (1989) 239–269] if and only if A is generated by the normalizing groupoid of B. Moreover, we show that there exists the largest intermediate equivalence subrelation which contains as a normal subrelation. We further give a definition of “commensurability groupoid” as a generalization of normality. We show that the commensurability groupoid of B in A generates A if and only if the inclusion BA is discrete in the sense of Izumi–Longo–Popa [M. Izumi, R. Longo, S. Popa, A Galois correspondence for compact groups of automorphisms of von Neumann algebras with a generalization to Kac algebras, J. Funct. Anal. 155 (1998) 25–63]. We also show that there exists the largest equivalence subrelation such that the inclusion is discrete. It turns out that the intermediate equivalence subrelations and thus defined can be viewed as groupoid-theoretic counterparts of a normalizer subgroup and a commensurability subgroup in group theory. 相似文献
4.
Let G be a unipotent algebraic subgroup of some defined over . We describe an algorithm for finding a finite set of generators of the subgroup . This is based on a new proof of the result (in more general form due to Borel and Harish-Chandra) that such a finite generating set exists. 相似文献
5.
We prove an index theorem for boundary value problems in Boutet de Monvel's calculus on a compact manifold X with boundary. The basic tool is the tangent semi-groupoid generalizing the tangent groupoid defined by Connes in the boundaryless case, and an associated continuous field of C*-algebras over [0,1]. Its fiber in =0, , can be identified with the symbol algebra for Boutet de Monvel's calculus; for ≠0 the fibers are isomorphic to the algebra of compact operators. We therefore obtain a natural map . Using deformation theory we show that this is the analytic index map. On the other hand, using ideas from noncommutative geometry, we construct the topological index map and prove that it coincides with the analytic index map. 相似文献
6.
A discrete time invariant linear state/signal system Σ with a Hilbert state space and a Kren signal space has trajectories (x(),w()) that are solutions of the equation , where F is a bounded linear operator from into with a closed domain whose projection onto is all of . This system is passive if the graph of F is a maximal nonnegative subspace of the Kren space . The future behavior of a passive system Σ is the set of all signal components w() of trajectories (x(),w()) of Σ on with x(0)=0 and . This is always a maximal nonnegative shift-invariant subspace of the Kren space , i.e., the space endowed with the indefinite inner product inherited from . Subspaces of with this property are called passive future behaviors. In this work we study passive state/signal systems and passive behaviors (future, full, and past). In particular, we define and study the input and output maps of a passive state/signal system, and the past/future map of a passive behavior. We then turn to the inverse problem, and construct two passive state/signal realizations of a given passive future behavior , one of which is observable and backward conservative, and the other controllable and forward conservative. Both of these are canonical in the sense that they are uniquely determined by the given data , in contrast earlier realizations that depend not only on , but also on some arbitrarily chosen fundamental decomposition of the signal space . From our canonical realizations we are able to recover the two standard de Branges–Rovnyak input/state/output shift realizations of a given operator-valued Schur function in the unit disk. 相似文献
7.
Let M be a smooth, compact, orientable, weakly pseudoconvex manifold of dimension 3, embedded in (N2), of codimension one or more, and endowed with the induced CR structure. Assuming that the tangential Cauchy-Riemann operator has closed range in L2(M) in order to rule out the Rossi example, we push regularity up to show has closed range in Hs(M) for all s>0. We then use the Szegö projection to show there is a smooth solution for the problem given smooth data. The results are obtained via microlocalization by piecing together estimates for functions and (0,1) forms that hold on different microlocal regions. 相似文献
8.
J. Fernndez-Ochoa J. Martínez-Moreno J.M. Quesada 《Journal of Approximation Theory》2006,140(2):147-153
Let hp, 1<p<∞, be the best ℓp-approximation of the element from a proper affine subspace K of , hK, and let denote the strict uniform approximation of h from K. We prove that there are a vector and a real number a, 0a1, such thatfor all p>1, where with γp=o(ap/p). 相似文献
9.
Mosco convergence of Dirichlet forms in infinite dimensions with changing reference measures 总被引:1,自引:0,他引:1
Alexander V. Kolesnikov 《Journal of Functional Analysis》2006,230(2):382-418
Let E be an infinite-dimensional locally convex space, let {μn} be a weakly convergent sequence of probability measures on E, and let be a sequence of Dirichlet forms on E such that is defined on L2(μn). General sufficient conditions for Mosco convergence of the gradient Dirichlet forms are obtained. Applications to Gibbs states on a lattice and to the Gaussian case are given. Weak convergence of the associated processes is discussed. 相似文献
10.
Let , with
-1=x0n<x1n<<xnn<xn+1,n=1