Fourier analysis for vector-measures on compact abelian groups (Q326710)

The aim of this interesting paper is to analyze the properties of the extension of the fundamental tools of Fourier analysis on compact groups to the case of vector measures defined in these spaces. Following the research initiated in [\textit{O. Delgado} and \textit{P. J. Miana}, J. Math. Anal. Appl. 358, No. 2, 355--363 (2009; Zbl 1182.46035)] and [\textit{J. M. Calabuig} et al., J. Fourier Anal. Appl. 19, No. 2, 312--332 (2013; Zbl 1319.46031)], the author extends some results and answers some fundamental questions that appear in a natural way when this vector-valued generalization is considered. In particular, the Fourier transform of vector measures and convolution between scalar and vector-valued regular measures on the Borel sets of a compact abelian group are analyzed. If \(f\) is an integrable function with respect to the vector measure \(\nu\) defined on the compact abelian group \(G\), its Fourier transform is given by the function acting in the dual group \(\Gamma\) \[ \hat f^\nu(\gamma)= \int_G f(t) \, \overline{\gamma(t)} \, d \nu(t), \quad \gamma \in \Gamma. \] Among other relevant facts, some results on the subclasses of vector measures for which \(\hat f^\nu \in c_0(\Gamma,X)\) holds are given, showing in this way in which cases this extension of the Riemann-Lebesgue lemma holds. Indeed, under the assumption that the Banach space \(X\) satisfies a new version of the Radon-Nikodým property, we can assert that it has the Riemann-Lebesgue property, which concerns the set of all vector measures with values in \(X\). Once the question is solved for vector measures, it can be transferred to operators just by considering the vector measure defined by a given operator. Section 4 is devoted to analyze convolution between a scalar measure and a vector-valued one. It is shown that this convolution commutes and the following version of the classical formula for the Fourier transform holds, \[ \widehat{ \mu \ast \nu} ( \gamma)= \hat{\mu}(\gamma) \, \hat{\nu}(\gamma), \quad \gamma \in \Gamma, \] where \(\mu\) is a scalar measure and \(\nu\) a vector measure of a certain class. Section 5 provides some results on invariance under homomorphisms and translation invariant measures. Finally, Section 6 provides some applications, improving the results given in [loc. cit.] for convolution of integrable functions with respect to a vector measure.