Noise sequences of infinite matrices and their applications to the characterization of the.pdfVIP

a r X i v : q u a n t - p h / 0 6 0 1 1 5 5 v 1 2 3 J a n 2 0 0 6 NOISE SEQUENCES OF INFINITE MATRICES AND THEIR APPLICATIONS TO THE CHARACTERIZATION OF THE CANONICAL PHASE AND BOX LOCALIZATION OBSERVABLES PEKKA LAHTI, MACIEJ J. MACZYNSKI, EGON SCHEFFOLD, AND KARI YLINEN Abstract. Noise sequences of infinite matrices associated with covariant phase and box localization observables are defined and determined. The canonical observables are characterized within the relevant classes of observables as those with asymp- totically minimal or minimal noise, i.e., the noise tending to 0 or having the value 0. 1. Introduction In this paper we study infinite matrices whose entries are complex numbers of abso- lute value at most 1 and which are indexed by the sets N2 and Z2. In other words, we consider functions A : N × N → D and B : Z × Z → D, where D = {z ∈ C | |z| ≤ 1}. The sets FN D and FZ D of such functions are convex in a natural way and their extremal elements are the torus valued functions A : N × N → T and B : Z × Z → T, respec- tively. For each matrix A ∈ FN D , resp. B ∈ FZ D , we define a probabilistic concept called noise sequence (Sect. 2, resp. Sect. 4) which is used to characterize the extremal set FN T , resp. FZ T (Propositions 1 and 2, respectively). A function A : N×N → D is positive semidefinite if for each sequence (cn)n∈N ? C, for which cn 6= 0 for only finitely many n ∈ N, ∑ n,m cnA(n,m)cm ≥ 0. Such a function is normalized if A(n, n) = 1 for all n ∈ N. Positive semidefinite normalized matrices A ∈ FN D have an important application in quantum mechanics as they characterize the N-covariant semispectral measures, known also as the phase shift covariant phase observables. Among them are the so-called canonical phase observable, associated with the constant one function, and its unitary equivalents (in the sense of covariance systems) which are associated with the T-valued functions. Proposition 1 of Sec. 2 leads in Corollary 1 to a characterization o