Integral extension of multiplier operators in \(A(G)\) (Q365171): Difference between revisions

Assume throughout this review that \(G\) is an infinite, compact abelian group with normalized Haar measure \(\mu\) and dual group \(\Gamma\). Recall that the Banach algebra \(A(G)\) consists of all functions \(f\) in \(L^1(G)\) for which the Fourier transform \(\widehat{f}\) of \(f\) belongs to \(\ell^1 (\Gamma)\). For each \(\varphi \in \ell^{\infty}(\Gamma)\) a continuous, translation invariant linear operator \(T_{\varphi}\) can be defined from \(A(G)\) to \(A(G)\) by \((T_{\varphi}(f))^{\widehat{}} = \varphi \widehat{f}\). In other words, \(T_{\varphi}\) is a multiplication operator on \(A(G)\). Moreover, these are the only multiplication operators on \(A(G)\), that is the algebra \(\mathcal{M}(A(G))\) of all multiplier operators on \(A(G)\) can be identified with \(\ell^{\infty}(\Gamma)\). In Section 2 of the paper under review the authors characterize several operator ideals of \(\mathcal{M}(A(G))\) in terms of subspaces of \(\ell^{\infty}(\Gamma)\). Specifically, they show \(T_{\varphi}\) is a compact operator if and only if \(\varphi \in c_0(\Gamma)\); \(T_{\varphi}\) is nuclear if and only if \(\varphi \in \ell^1(\Gamma)\); and \(T_{\varphi}\) is 1-summing is equivalent to \(\varphi \in \ell^2(\Gamma)\). Let \(\varphi \in \ell^2(\Gamma)\) and let \(\mathcal{B}(G)\) be the Borel \(\sigma\)-algebra on \(G\). The optimal domain \(L^1(m_{\varphi})\) is studied in Section 3 of the paper. The space \(L^1(m_{\varphi})\) is optimal in the sense that it is the largest \(\sigma\)-order Banach function space over \((G, \mathcal{B}(G), \mu)\) which contains \(A(G)\) and for which \(T_{\varphi} : A(G) \rightarrow A(G)\) has a continuous \(A(G)\)-valued linear extension. We now describe \(L^1(m_{\varphi})\). The set function from \(\mathcal{B}(G)\) to \(A(G)\) defined by \[ m_{\varphi} : E \mapsto T_{\varphi} (\chi_E) = \check{\varphi} \ast \chi_E \] is actually a vector measure. Here \(\check{\varphi} \in L^2(G)\) is the inverse Fourier transform of \(\varphi\) and \(\chi_E\) is the characteristic function on \(E\). It is then shown that \[ L^1(m_{\varphi}) = \{ f \in L^1(G) : \check{\varphi} \ast (\chi_E f) \in A(G), \forall E \in \mathcal{B}(G) \}. \] Because \(\varphi \in \ell^2(\Gamma)\) the associated integration map \(I_{m_{\varphi}} : L^1(m_{\varphi} ) \rightarrow A(G)\) is given by \[ I_{m_{\varphi}} (f) = \check{\varphi} \ast f, \,\, f \in L^1(m_{\varphi}). \] The map \(I_{m_{\varphi}}\) is a continuous linear extension of \(T_{\varphi}\) from \(A(G)\) to \(L^1(m_{\varphi})\). It is also shown in Section 3 that if \(\varphi \in \ell^2(\Gamma)\) then we have the following containments \[ A(G) @>{\scriptstyle \subset}>>{\scriptstyle \not=} C(G) @>{\scriptstyle \subset}>>{\scriptstyle \not=} L^{\infty}(G) @>{\scriptstyle \subset}>>{\scriptstyle \not=} L^2(G) @>{\scriptstyle \subset}>>{\scriptstyle \not=} L^1(m_{\varphi}) \subseteq L^1(G) \] and that the last containment is proper if \(\varphi \in \ell^2(\Gamma) \setminus \ell^1(\Gamma)\). In the last section of the paper the authors show that for \(\varphi \in \ell^2(\Gamma), L^1(m_{\varphi})\) is a commutative Banach algebra. After establishing this fact several Banach algebra properties of \(L^1(m_{\varphi})\) are investigated.