Multisummability in Carleman ultraholomorphic classes by means of nonzero proximate orders (Q2414772)

The aim of this long paper is to to put forward the corresponding multisummability theory, in Balser's sense, of formal power series for finitely many levels determined by pairwise comparable, nonequivalent weight sequences admitting nonzero proximate orders. They extend the powerful multisummability theory for finitely many Gevrey levels, developed by Ramis, Écalle, Balser, Sibuya, Martinet, Malgrange and Braaksma, among others. The departure point of the authors is the summability theory in ultraholomorphic classes defined in terms of a weight sequence admitting a nonzero proximate order, developed by \textit{A. Lastra} et al. [J. Math. Anal. Appl. 430, No. 2, 1175--1206 (2015; Zbl 1327.30004)], which generalized the Gevrey \(k\)-summability of Ramis from the late 1970's, by means of the moment methods of \textit{W. Balser} [Formal power series and linear systems of meromorphic ordinary differential equations. New York, NY: Springer (2000; Zbl 0942.34004)]. The main result of the first part of the paper is Watson's lemma, characterizing the injectivity of the Borel map. Proximate orders and the theory of regular variation for functions and sequences are treated later. The construction and the properties of the kernels for \(M\)-summability and the corresponding, formal and analytic, Laplace and Borel transforms are recalled, allowing the authors to explicitly obtain the \(M\)-sums of formal power series \(M\)-summable in a direction. In Section 3, after establishing the basic properties of the quotient and product sequences of two weight sequences, the Tauberian theorem is obtained, which allows for a consistent definition of multisummability whenever the growth indices of the sequences involved are mutually distinct. Section 4 contains a purely analytical approach to multisummability. The authors first prove that a kernel of summability is uniquely determined by its sequence of moments. Strong kernels of summability are introduced in order to obtain natural summability kernels for the quotient and product sequences of two sequences, so giving rise, respectively, to convolution and acceleration kernels and operators, as developed by Balser [loc. cit.]. Later, they present a procedure for the explicit reconstruction, through acceleration, of the multisum of a multisummable series. The last part of the paper contains a cohomological approach to multisummability, following the one established by \textit{B. Malgrange} and \textit{J.-P. Ramis} [Ann. Inst. Fourier 42, No. 1--2, 353--368 (1992; Zbl 0759.34007)], so focusing on the algebraic framework of the theory.