Extension of Pettis integration: Pettis operators and their integrals (Q2311335)

Due to some well-known structure problems, spaces of (vector-valued measurable) Pettis integrable functions \(P_1(\mu,X)\) have some deficits when compared with other usual function spaces. In this paper, and with the aim of providing an adapted definition that allows to overcome -- or at least, to understand in a more general context -- the problems regarding these spaces, the authors propose a new context. In it, integrals are substituted by weak\(*\)-weak operators from \(X^*\) to \(L^1(\mu)\), which they call Pettis operators, ``as objects that are subject to integration,'' as explained in the paper. The motivation of this approach is made using the equivalence of Pettis integrability with the existence of certain (countably additive) vector measures, passing through Bartle-Dunford-Schwartz integration of scalar-valued functions with respect to vector measures. Several issues are addressed from this point of view. For example, the space of Pettis operators is identified with certain spaces of vector measures and of vector-valued harmonic functions. Indeed, one of the main results establishes that it coincides with the space of \(\mu\)-continuous countably additive \(X\)-valued measures (Theorem~3.1). Also, characterizations of main properties of the vector measure associated to Pettis operators are given in terms of the order boundedness of its range (Theorem~3.3). Martingales associated to Pettis operators are analysed in Section~4, motivated also by the fact that some relevant properties regarding them related to Pettis integration fail (see Theorem~4.1), as well as Fubini-type theorems, for which some interesting results are provided in Section~5. Finally, Sections~6 and~7 show a nice association of Pettis-type integration (in its extended form of Pettis operators) with some classic results in harmonic analysis. Pettis operators are successfully connected with vector-valued harmonic functions, providing a creative and -- as far as I know -- completely new way of understanding the topic. This abstract link becomes a useful tool, as shown in Section~7, where the authors explain how to adapt the theory, showing in Theorem~7.3 a surprising association (surprising at least at first sight). The authors warn at the beginning of the article that they do not intend to develop a complete new theory on the topic they introduce. After reading the article, one may note that this is indeed true, but in my opinion it clearly opens the door to this new theory, which would be proposed to young mathematicians as a way to continue the research on this relevant topic with new ideas.