DOCSLIB.ORG

Approximate Identities in Banach Function Algebras H. G. Dales

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Approximate Identities in Banach Function Algebras H. G. Dales Approximate identities in Banach function algebras H. G. Dales, Lancaster National University of Ireland, Maynooth, 19 June 2013 In honour of the retirement of Anthony G. O'Farrell Work in progress with Ali Ulger,¨ Istanbul 1 Definitions Let K be a locally compact space. Then C0(K) is the space of all complex-valued, continuous functions on K that vanish at infinity. This is a commutative Banach algebra with respect to the uniform norm j · jK. A function algebra on K is a subalgebra of C0(K) such that, for x; y 2 K with x 6= y, there is f 2 A with f(x) 6= f(y), and, for each x 2 K, there is f 2 A with f(x) 6= 0. A Banach function algebra on K is a function algebra A that is a Banach algebra for a norm k · k, so that kfgk ≤ kfk kgk for f; g 2 A. Necessarily, kfk ≥ jfjK for f 2 A. The algebra A is a uniform algebra if it is closed in C0(K). 2 Natural Banach function algebras A Banach function algebra A on K is natural if every character on A has the form f 7! f(x) = "x(f) for some x 2 K. Equiv- alently, every maximal modular ideal has the form Mx = ff 2 A : f(x) = 0g for some x 2 K. Every commutative, semisimple Banach algebra is a Banach function algebra on its character space. 3 Approximate identities Let (A; k · k) be a natural Banach function algebra on K. An approximate identity is a net (eα) in A such that kfeα − fk ! 0 for each f 2 A. A pointwise approximate identity is a net (eα) in A such that eα ! 1 pointwise on K. Clearly, each approximate identity is a point- wise approximate identity. These approximate identities (eα) are bounded if sup keαk < 1 (and the sup is the bound); they are contractive if the bound is 1. Say BAI and BPAI, CAI and CPAI. The algebra A is (pointwise) contractive if every Mx has a C(P)AI. We are interested in the best bound of BAIs. 4 Factorization Let A be an algebra. Then A[2] = fab : a; b 2 Ag;A2 = lin A[2] : Cohen's factorization theorem Let A be a commutative Banach algebra. Then the fol- lowing are equivalent: (a) A has a BAI of bound m; (b) for each " > 0, each a 2 A can be written as a = bc, where kbk ≤ m and ka − ck < " (and so A = A[2], and A factors). 2 [Much more is true.] 5 Examples (I) Take K compact and A = C(K). Then every Mx has a CAI. Are there any more (pointwise) contractive uniform algebras? (II) Let G be a LCA group (e.g., G = Z or 1 G = R), and let A be the group algebra (L (G);?). The Fourier transform maps L1(G) onto the Fourier algebra A(Γ), where Γ is the dual group to G. Now A(Γ) is a natural Banach function algebra on Γ. The algebra A(Γ) always has a CAI, and all the maximal modular ideals Mγ have a BAI, of bound 2 (very standard); further, `2' is the best bound of a BAI in Mγ (less well-known; Derighetti). 6 More examples Look at natural Banach function algebras on I = [0; 1]. (III) Take A = Lip α(I), where α > 0, so that (jf(x) − f(y)j) kfk = jfj + sup α I jx − yjα for f 2 A. If f(0) = 0 and f(1=n) = 1, then α kfk ∼ n , so there are no BPAI in M0. (IV) Look at BV C(I), the algebra of continu- ous functions of bounded variation on I. Here kfk = jfj + var(f) var I for f 2 BV C(I). There is an obvious BAI in M0 of bound 2. 7 The Choquet boundary Let A be a Banach function algebra on K. A closed subset F of K is a peak set if there exists a function f 2 A with f(x) = 1 (x 2 F ) and jf(y)j < 1 (y 2 K n F ); in this case, f peaks on F ; a point x 2 K is a peak point if fxg is a peak set, and a p -point if fxg is an intersection of peak sets. The set of p -points of A is Γ0(A), the Choquet boundary of A. Its closure is the Silovˇ boundary. Theorem Let A be a Banach function algebra on K. Then Mx has a BAI ) x 2 Γ0(A). 2 Example Let A be the disc algebra on D. Then ff 2 A : f(0) = 0g does not have a BAI. This is obvious anyway. 2 8 Uniform algebras Theorem Let A be a uniform algebra on com- pact K. Then the following are equivalent: (a) x 2 Γ0(A); (b) Mx has a BAI; (c) "x is an extreme point of the state space; (d) Mx has a CAI. The implication (a) ) (d): Take f that peaks n at x. Then (1K − f : n 2 N) is a BAI of bound n 2 for Mx. One can modify the functions f into functions gn 2 A such that (1K − gn : n 2 N) is a CAI for Mx. 2 Definition A natural uniform algebra on a com- pact space K is a Cole algebra if Γ0(A) = K. Important fact There are Cole algebras other than C(K). 9 First question Question What is the relation between the existence of a CPAI and a CAI in Banach function algebras? Remark Suppose that there is a bounded net (eα) in A such that feα ! f weakly for each f 2 A. Then A has a BAI, with the same bound. Original example { Jones and Lahr, 1977 +• Let S = (Q ; +) be the semigroup of strictly positive rational numbers. The semigroup algebra A = (` 1(S);? ) is a natural Banach function algebra on Sb, the collection of semi- characters on S. Then there is a sequence (nd) in such that (δ ) is a CPAI for A. However, N 1=nd A does not have any approximate identity. This example is not pointwise contractive. 2 10 First question - more examples Example Let G be a LCA group that is not compact. Then the minimum bound of a BAI in a maximal ideal of A(Γ) is 2. We can show that the minimum bound of a BPAI in a maximal ideal is 1 p (1 + 2) > 1 ; 2 so L1(G) is not pointwise contractive. We do not know if this constant is best-possible. 2 Example Let G be a LCA group that is not discrete. Let n 1 1 o A = f 2 L (G): fb 2 L (Γ) ; with kfk = maxfkfk ; fb g. Then A is a nat- 1 1 ural Banach function algebra on Γ. It has a CPAI, but no BAI. 2 11 First question - more examples 1 Example Let B = (L (R);?). Then there is a singular measure µ0 on R such that µ0 ? µ0 2 B (Hewitt). Look at A = B ⊕ Cµ0 as a closed subalgebra of (M(R);?). Then A has CPAI, but no approximate identity at all because A2 = B, which is not dense in A. (Hence CPAI 6) factorization.) 2 Example Let A be a Banach function algebra that is reflexive as a Banach space. For exam- + ple, take the set of sequences α = (αn) on Z such that 0 1 11=2 X 2 2 kαk = @ jαnj (1 + n) A < 1 ; n=0 with convolution product, giving a Banach func- tion algebra on D. Suppose that Mx has a BPAI. Then A contains the characteristic function of K n fxg, and so x is isolated. Thus maximal ideals of our exam- ple do not have BPAIs. 2 12 Uniform algebras So far we have not given a pointwise contrac- tive Banach function algebra that is not con- tractive. Fact Let A be a natural uniform algebra on K. Suppose that A has a BPAI (respectively, CPAI) that is a sequence. Then A has a BAI (respectively, CAI). Proof Suppose that (fn) is bounded and fn ! 1 pointwise on K. Then fn ! 1 weakly by the dominated convergence theorem. An earlier remark now shows that A has a BAI, with the same bound. 2 13 Feinstein's example Example Joel Feinstein has an amazing exam- ple of a natural uniform algebra A on a com- pact, metrizable space K such that there is a point x0 2 K with Γ0(K) = K n fx0g. The construction is a modification of Cole's original construction. Take M = Mx0. Each finite F disjoint from x0 is a peak set, and so there exists fF 2 M that peaks on F . The net ffF g is a CPAI for M. All the other Mx have a CAI, so A is pointwise contractive. But M does not have a BAI because x0 is not a peak point, and so A is not contractive. (However M factors.) 2 14 Second question Question Let A be a pointwise contractive, natural Banach function algebra. Is A neces- sarily a uniform algebra? Let A be a contractive, natural Banach func- tion algebra. Is A necessarily a Cole algebra? 15 A special case Definition Let S be a non-empty set. A Ba- nach sequence algebra on S is a Banach function algebra A on S such that c00(S) ⊂ A. Theorem Let A be a pointwise contractive, natural Banach sequence algebra on a set S.
Load more

Recommended publications
  • Weak and Norm Approximate Identities Are Different
    Pacific Journal of Mathematics WEAK AND NORM APPROXIMATE IDENTITIES ARE DIFFERENT CHARLES ALLEN JONES AND CHARLES DWIGHT LAHR Vol. 72, No. 1 January 1977 PACIFIC JOURNAL OF MATHEMATICS Vol. 72, No. 1, 1977 WEAK AND NORM APPROXIMATE IDENTITIES ARE DIFFERENT CHARLES A. JONES AND CHARLES D. LAHR An example is given of a convolution measure algebra which has a bounded weak approximate identity, but no norm approximate identity. 1* Introduction* Let A be a commutative Banach algebra, Ar the dual space of A, and ΔA the maximal ideal space of A. A weak approximate identity for A is a net {e(X):\eΛ} in A such that χ(e(λ)α) > χ(α) for all αei, χ e A A. A norm approximate identity for A is a net {e(λ):λeΛ} in A such that ||β(λ)α-α|| >0 for all aeA. A net {e(X):\eΛ} in A is bounded and of norm M if there exists a positive number M such that ||e(λ)|| ^ M for all XeΛ. It is well known that if A has a bounded weak approximate identity for which /(e(λ)α) —>/(α) for all feA' and αeA, then A has a bounded norm approximate identity [1, Proposition 4, page 58]. However, the situation is different if weak convergence is with re- spect to ΔA and not A'. An example is given in § 2 of a Banach algebra A which has a weak approximate identity, but does not have a norm approximate identity. This algebra provides a coun- terexample to a theorem of J.
    [Show full text]
  • On the Beurling Algebras A+ Α(D)—Derivations And
    IJMMS 2004:32, 1679–1701 PII. S0161171204309270 http://ijmms.hindawi.com © Hindawi Publishing Corp. ON THE BEURLING ALGEBRAS + D —DERIVATIONS Aα( ) AND EXTENSIONS HOLGER STEINIGER Received 27 September 2003 + D Based on a description of the squares of cofinite primary ideals of Aα( ), we prove the ≥ + D following results: for α 1, there exists a derivation from Aα( ) into a finite-dimensional module such that this derivation is unbounded on every dense subalgebra; for m ∈ N and ∈ + + D α [m,m 1), every finite-dimensional extension of Aα( ) splits algebraically if and only if α ≥ m+1/2. 2000 Mathematics Subject Classification: 46J15, 46M20, 46H40. 1. Introduction. Let α be a positive real number. By D, we denote the open unit + D D disk. The Beurling algebra Aα( ) is a subalgebra of the classical disk algebra A( ). ∈ D = ∞ n ∈ D For f A( ) with power series expansion f(z) n=0 anz (z ), the function f + D ∞ | | + α ∞ = belongs to Aα( ) if and only if n=0 an (n 1) < . In this case, we define f α : ∞ | | + α + D n=0 an (n 1) . Clearly, Aα( ) is a Banach algebra with respect to this norm. These algebras have been considered in [13] where results on primary ideals were applied to operator theory. More recently, the algebras have appeared in the examina- tion of finite-dimensional extensions of a whole range of commutative Banach algebras + D [4]. The present paper deals with continuity problems of derivations from Aα( ) and with finite-dimensional extensions of this special type of Beurling algebras. Some of the results of the first paper will be the starting point for our investigation.
    [Show full text]
  • AUTOMATIC CONTINUITY of CERTAIN ISOMORPHISMS BETWEEN REGULAR BANACH FUNCTION ALGEBRAS by JUAN J
    AUTOMATIC CONTINUITY OF CERTAIN ISOMORPHISMS BETWEEN REGULAR BANACH FUNCTION ALGEBRAS by JUAN J. FONTt (Received 9 May, 1996) Abstract. Let A and B be regular semisimple commutative Banach algebras; that is to say, regular Banach function algebras. A linear map T denned from A into B is said to be separating or disjointness preserving if/.g = 0 implies Tf.Tg = 0, for all f,g e A In this paper we prove that if A satisfies Ditkin's condition then a separating bijection is automatically continuous and its inverse is separating. If also B satisfies Ditkin's condition, then it induces a homeomorphism between the structure spaces of A and B. Finally, we show that linear isometries between regular uniform algebras are separating. As corollaries, a classical theorem of Nagasawa ([19]) and the Banach-Stone theorem (both for regular uniform algebras) are easily inferred. 1. Introduction. The basic problem in the theory of automatic continuity of linear operators consists of giving algebraic conditions on two Banach algebras si and 38 which ensure that every algebra homomorphism !T:s4-*2ft is necessarily continuous. The most important positive result in this context is due to B. E. Johnson ([15]). Every algebra homomorphism of a Banach algebra onto a semisimple Banach algebra is continuous. In this paper we shall study the automatic continuity of a special type of linear mapping which extends the concept of algebra homomorphism: Let A and B be two Banach algebras. Then a linear map T denned from A into B is said to be separating or disjointness preserving (also called d-homomorphisms, Lamperti operators,...) if f.g — 0 implies Tf.Tg=0, for all/,g e A.
    [Show full text]
  • Regularity Conditions for Banach Function Algebras
    Regularity conditions for Banach function algebras Dr J. F. Feinstein University of Nottingham June 2009 1 1 Useful sources A very useful text for the material in this mini-course is the book Banach Algebras and Automatic Continuity by H. Garth Dales, London Mathematical Society Monographs, New Series, Volume 24, The Clarendon Press, Oxford, 2000. In particular, many of the examples and conditions discussed here may be found in Chapter 4 of that book. We shall refer to this book throughout as the book of Dales. Most of my e-prints are available from www.maths.nott.ac.uk/personal/jff/Papers Several of my research and teaching presentations are available from www.maths.nott.ac.uk/personal/jff/Beamer 2 2 Introduction to normed algebras and Banach algebras 2.1 Some problems to think about Those who have seen much of this introductory material before may wish to think about some of the following problems. We shall return to these problems at suitable points in this course. Problem 2.1.1 (Easy using standard theory!) It is standard that the set of all rational functions (quotients of polynomials) with complex coefficients is a field: this is a special case of the “field of fractions" of an integral domain. Question: Is there an algebra norm on this field (regarded as an algebra over C)? 3 Problem 2.1.2 (Very hard!) Does there exist a pair of sequences (λn), (an) of non-zero complex numbers such that (i) no two of the an are equal, P1 (ii) n=1 jλnj < 1, (iii) janj < 2 for all n 2 N, and yet, (iv) for all z 2 C, 1 X λn exp (anz) = 0? n=1 Gap to fill in 4 Problem 2.1.3 Denote by C[0; 1] the \trivial" uniform algebra of all continuous, complex-valued functions on [0; 1].
    [Show full text]
  • Operator Algebras with Contractive Approximate Identities
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Elsevier - Publisher Connector Journal of Functional Analysis 261 (2011) 188–217 www.elsevier.com/locate/jfa Operator algebras with contractive approximate identities David P. Blecher a,∗,1, Charles John Read b a Department of Mathematics, University of Houston, Houston, TX 77204-3008, United States b Department of Pure Mathematics, University of Leeds, Leeds LS2 9JT, England, United Kingdom Received 15 November 2010; accepted 24 February 2011 Available online 21 March 2011 Communicated by S. Vaes Abstract We give several applications of a recent theorem of the second author, which solved a conjecture of the first author with Hay and Neal, concerning contractive approximate identities; and another of Hay from the theory of noncommutative peak sets, thereby putting the latter theory on a much firmer foundation. From this theorem it emerges there is a surprising amount of positivity present in any operator algebras with contractive approximate identity. We exploit this to generalize several results previously available only for ∗ C -algebras, and we give many other applications. © 2011 Elsevier Inc. All rights reserved. Keywords: Operator algebras; One-sided ideals; Hereditary subalgebra; Approximate identity; Peak set; Pseudo-invertible elements; Completely positive operator 1. Introduction An operator algebra is a closed subalgebra of B(H), for a Hilbert space H . We recall that by a theorem due to Ralf Meyer, every operator algebra A has a unique unitization A1 (see [30] or [10, Section 2.1]). Below 1 always refers to the identity of A1 if A has no identity.
    [Show full text]
  • 1. Introduction
    FACTORIZATION IN COMMUTATIVE BANACH ALGEBRAS H. G. DALES, J. F. FEINSTEIN, AND H. L. PHAM Abstract. Let A be a (non-unital) commutative Banach algebra. We consider when A has a variety of factorization properties: we list the (ob- vious) implications between these properties, and then consider whether any of these implications can be reversed in various classes of commu- tative Banach algebras. We summarize the known counter-examples to these possible reverse implications, and add further counter-examples. Some results are used to show the existence of a large family of prime ideals in each non-zero, commutative, radical Banach algebra with a dense set of products. 1. Introduction Let A be a (non-unital) commutative Banach algebra. We wish to examine when A factors in a variety of senses. Our main results are counter-examples to a number of questions that have been raised. Indeed, we shall list seven such factorization properties, called (I)(VII), and note that each of these immediately implies the next one. We shall also, in x4, discuss two other `lo- cal' factorization properties, called (A) and (B) (where (A) ) (B)); these properties are relevant for Esterle's classication of commutative, radical Banach algebras that is given in [14]. We shall then discuss whether or not any of these implications can be reversed when we restrict attention to par- ticular classes of commutative Banach algebras. We shall show that several cannot be reversed, but we leave open other possible reverse implications. A summary in x6 describes our knowledge at the present time. We shall concentrate on two particular classes of commutative Banach algebras A: rst, on the case where A is semi-simple (so that A is a Banach function algebra), and in particular when A is a maximal ideal in a uniform algebra on a compact space, and, second, on the other extreme case where A 2010 Mathematics Subject Classication.
    [Show full text]
  • Banach Function Algebras with Dense Invertible Group
    PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 136, Number 4, April 2008, Pages 1295–1304 S 0002-9939(07)09044-2 Article electronically published on December 21, 2007 BANACH FUNCTION ALGEBRAS WITH DENSE INVERTIBLE GROUP H. G. DALES AND J. F. FEINSTEIN (Communicated by N. Tomczak-Jaegermann) Abstract. In 2003 Dawson and Feinstein asked whether or not a Banach function algebra with dense invertible group can have a proper Shilov bound- ary. We give an example of a uniform algebra showing that this can happen, and investigate the properties of such algebras. We make some remarks on the topological stable rank of commutative, unital Banach algebras. In particular, we prove that tsr(A) ≥ tsr(C(ΦA)) whenever A is approximately regular. 1. Introduction Let A be a commutative, unital Banach algebra. The character space of A is denoted by ΦA,asin[8].Fora ∈ A, we denote the Gel’fand transform of a by a.WesaythatΦA contains analytic structure if there is a continuous injection τ from the open unit disk D to ΦA such that, for all a ∈ A, a ◦ τ is analytic on D. In [16], Stolzenberg gave a counter-example to the conjecture that, whenever a uniform algebra has proper Shilov boundary, its character space must contain analytic structure (see also [17, Theorem 29.19], [19] and [1]). Cole gave an even more extreme example in his thesis [3], where the Shilov boundary is proper and yet every Gleason part is trivial. It is elementary to show that the invertible group of A cannot be dense in A whenever ΦA contains analytic structure.
    [Show full text]
  • A Covariance Function Approach to Prior Specification for Bayesian
    Journal of Machine Learning Research 22 (2021) 1-57 Submitted 11/20; Revised 5/21; Published 6/21 The Ridgelet Prior: A Covariance Function Approach to Prior Specification for Bayesian Neural Networks Takuo Matsubara [email protected] School of Mathematics, Statistics & Physics Newcastle University Newcastle upon Tyne, NE1 7RU, United Kingdom Chris J. Oates [email protected] School of Mathematics, Statistics & Physics Newcastle University Newcastle upon Tyne, NE1 7RU, United Kingdom Fran¸cois-Xavier Briol [email protected] Department of Statistical Science University College London London, WC1E 6BT, United Kingdom Editor: Pierre Alquier Abstract Bayesian neural networks attempt to combine the strong predictive performance of neural networks with formal quantification of uncertainty associated with the predictive output in the Bayesian framework. However, it remains unclear how to endow the parameters of the network with a prior distribution that is meaningful when lifted into the output space of the network. A possible solution is proposed that enables the user to posit an appropriate Gaussian process covariance function for the task at hand. Our approach constructs a prior distribution for the parameters of the network, called a ridgelet prior, that approximates the posited Gaussian process in the output space of the network. In contrast to existing work on the connection between neural networks and Gaussian processes, our analysis is non-asymptotic, with finite sample-size error bounds provided. This establishes the universality property that a Bayesian neural network can approximate any Gaussian process whose covariance function is sufficiently regular. Our experimental assessment is limited to a proof-of-concept, where we demonstrate that the ridgelet prior can out-perform an unstructured prior on regression problems for which a suitable Gaussian process prior can be provided.
    [Show full text]
  • Categorical Characterizations of Operator-Valued Measures
    Categorical characterizations of operator-valued measures Frank Roumen Inst. for Mathematics, Astrophysics and Particle Physics (IMAPP) Radboud University Nijmegen [email protected] The most general type of measurement in quantum physics is modeled by a positive operator-valued measure (POVM). Mathematically, a POVM is a generalization of a measure, whose values are not real numbers, but positive operators on a Hilbert space. POVMs can equivalently be viewed as maps between effect algebras or as maps between algebras for the Giry monad. We will show that this equivalence is an instance of a duality between two categories. In the special case of continuous POVMs, we obtain two equivalent representations in terms of morphisms between von Neumann algebras. 1 Introduction The logic governing quantum measurements differs from classical logic, and it is still unknown which mathematical structure is the best description of quantum logic. The first attempt for such a logic was discussed in the famous paper [2], in which Birkhoff and von Neumann propose to use the orthomod- ular lattice of projections on a Hilbert space. However, this approach has been criticized for its lack of generality, see for instance [22] for an overview of experiments that do not fit in the Birkhoff-von Neumann scheme. The operational approach to quantum physics generalizes the approach based on pro- jective measurements. In this approach, all measurements should be formulated in terms of the outcome statistics of experiments. Thus the logical and probabilistic aspects of quantum mechanics are combined into a unified description. The basic concept of operational quantum mechanics is an effect on a Hilbert space, which is a positive operator lying below the identity.
    [Show full text]
  • NOTES and REMARKS Chapter I.I. the Concept of a Bounded
    NOTES AND REMARKS Chapter I.i. The concept of a bounded approximate identity goes back to the earliest studies of the group algebra LI(G) [282]. Approximate identities in LI(G) are discussed in a systematic way in A. Weil's book [281], especially pp. 52, 79-80, and 85-86, and they have since become a standard tool in harmonic analysis [131], §28 (called bounded approximate unit); [80], Chapter III (called approximate identity); [147], Chapter I (called summability kernel); [220], §6 (called bounded multiple units). For applications to group representations it soon became necessary to consider C*-algebras without identity element. In order to discuss such algebras I. E. Segal constructed for every C*-algebra an approximate identity bounded by one [243], Lemma i.i. Many results about a C*-algebra without an identity element can be obtained by embedding such an algebra in a C*-algebra with identity element, called adJunction of an identity (see [222], Lemma (4.1.13) or [70], 1.3.8). But some problems, especially those which involve approximate identities, are not susceptible to this approach. Therefore, the approximate identity is the main tool in Dixmier's book [70] (called unite approchee in the earlier French editions) to carry through all the basic theory of C*-algebras. In recent years many results known for C*-algebras or for Banach algebras with identity have been extended to algebras with an approximate identity. The mainspring for much of this work was the Cohen-Hewitt 238 NOTES AND REMARKS factorization theorem for Banach modules [131], §32. The elementary results in I.l are very familiar.
    [Show full text]
  • AMENABILITY PROPERTIES of BANACH ALGEBRA of BANACH ALGEBRA-VALUED CONTINUOUS FUNCTIONS by Reza Ghamarshoushtari a Thesis Submitt
    AMENABILITY PROPERTIES OF BANACH ALGEBRA OF BANACH ALGEBRA-VALUED CONTINUOUS FUNCTIONS by Reza Ghamarshoushtari A Thesis submitted to the Faculty of Graduate Studies of The University of Manitoba in partial fulfilment of the requirements for the degree of MASTER OF SCIENCE Department of Mathematics University of Manitoba Winnipeg Copyright c 2014 by Reza Ghamarshoushtari Abstract In this thesis we discuss amenability properties of the Banach algebra-valued continuous functions on a compact Hausdorff space X. Let A be a Banach algebra. The space of A- valued continuous functions on X, denoted by C(X; A), form a new Banach algebra. We show that C(X; A) has a bounded approximate diagonal (i.e. it is amenable) if and only if A has a bounded approximate diagonal. We also show that if A has a compactly central approximate diagonal then C(X; A) has a compact approximate diagonal. We note that, unlike C(X), in general C(X; A) is not a C∗-algebra, and is no longer commutative if A is not so. Our method is inspired by a work of M. Abtahi and Y. Zhang. In addition to the above investigation, we directly construct a bounded approximate diagonal for C0(X), the Banach algebra of the closure of compactly supported continuous functions on a locally compact Hausdorff space X. ii Acknowledgment I would like to thank my supervisor Dr. Yong Zhang for superlative and superior super- vision. His direction, inspiration, and endless enthusiasm about the subject have made my work much more interesting. I would like to thank him for his patience, understanding, believing that I can self-manage, for his help throughout my studentship, and for providing invaluable expertise and discussions with regard to matters relating to my subject.
    [Show full text]
  • ON the OPEN BALL CENTERED at an INVERTIBLE ELEMENT of a BANACH ALGEBRA 1. Introduction Let a Be a Complex Unital Banach Algebra
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Research Archive of Indian Institute of Technology Hyderabad O perators a nd & M atrices www.ele-math.com ON THE OPEN BALL CENTERED AT AN INVERTIBLE ELEMENT OF A BANACH ALGEBRA GEETHIKA SEBASTIAN AND SUKUMAR DANIEL Submitted to Operators and Matrices Abstract. Let A be a complex unital Banach algebra. Since the set of invertible elements is open, there is an open ball around every invertible element. In this article, we investigate the Banach algebras for which the radius given by the Neumann series is optimal. 1. Introduction Let A be a complex unital Banach algebra with unit e. G(A) and Sing(A) denote the set of invertible and singular elements of A respectively. The spectrum, spectral radius and the resolvent of an element a in A, are denoted by s(a), r(a) and r(a) respectively. It is well known that G(A) is open in A, as the open ball centered at any invert- a 1 B a; 1 G(A) ible element with radius ka−1k , denoted by ka−1k , is contained in ([2] Theorem 2.11). It is a natural to ask if there is a bigger open ball centered at a inside G(A)? It will be convenient to have the following definition. DEFINITION 1. An element a 2 G(A) is said to satisfy condition (B) (or belongs to the B class) if 1 B a; \ Sing(A) 6= f: ka−1k We say a Banach algebra A satisfies condition (B) if every a 2 G(A) satisfies condition (B).
    [Show full text]