\({\mathfrak p}\)-adic pseudodifferential calculus (Q2388665)

This extensive paper is a summary of the Ph.D. Thesis of the author [``Calcul pseudo-différentiel \(p\)-adique'', University of Reims (2001)]. The purpose is to develop the pseudodifferential analysis of operators acting on complex-valued functions on \(K^n\), where \(K\) is a non-archimedean valued local field. The article is an interesting contribution to \(p\)-adic functional analysis. One has just to have in mind the important applications that the classical Weyl calculus has in the study of partial differential equations. I encourage the author to investigate the possible applications of his paper in the non-archimedean setting. In the first half of the work, he calculates the Wigner function \(W(u,v)\) associated to two coherent states \(u,v \in L^2(K^n)\). This function is the symbol of the operator of range 1, \(w \mapsto \langle w,v\rangle u\). The calculation is used to deduce expressions, in terms of Wigner functions of coherent states, of the symbol of a sequentially continuous operator from the space \({\mathcal S}(K^n)\) of rapidly decreasing functions into the dual space \({\mathcal S}'(K^n)\) of tempered distributions. These expressions have also consequences on the regularity of operators. The first half of the article finishes with applications of the previous results to obtain the non-archimedean analogue of the characterization, given by \textit{R. Beals} [J. Math. Anal. Appl. 58, 32--49 (1977; Zbl 0353.35072)], of pseudodifferential operators \(A\) of classical analysis in terms of continuity properties of the iterated brackets of \(A\) with families of certain classes of operators. This characterization leads to the existence of a non-archimedean functional calculus of operators of weight 1. The main goal of the second half of the paper is to obtain a composition formula of the \(p\)-adic pseudodifferential calculus in one variable, that is, to examine the structure of the symbol \(f \# g\) of the composition of two Hilbert-Schmidt operators of symbols \(f\) and \(g\). After quite a bit of hard work, which involves a sophisticated machinery (e.g., the Mellin transformation, the Gamma factors, the decomposition of every function \(f \in L^2(K^2)\) as an integral superposition of homogeneous functions), the author finally gets the formula, which is completely different from the classical one. The reason of this difference is the following: since no derivation operators are available in \(p\)-adic analysis (apart from the polynomial functions with complex values), no Moyal-style expansion of the composition \(f \# g\) of two symbols \(f,g\) is possible. Nevertheless, using the theory of multiplicative characters of \(K^{\times}\), the author gives a composition formula expressing the decomposition in ``homogeneous terms'' of a sharp-product \(f \# g\) in terms of the corresponding decompositions in ``plane waves'' of each of the two factors. Several serious convergence difficulties appear (and are solved!) in order to have a \(p\)-adic harmonic oscillator, a key point to arrive at the desired composition formula. To carry out the above objectives, the author uses basic central tools of functional analysis, such as the notions of compact operator and nuclear space and the well-known Hahn-Banach theorem. The way he uses these tools requires the assumption of local compactness on \(K\) imposed by him. Non-locally compact ground fields are not considered in the paper. For these fields, the situation changes strongly and becomes typically non-archimedean. In fact, the only convex compact sets are the singletons. Also, for non-spherically complete fields, the Hahn-Banach theorem may fail. I think that it would be interesting for a future work to investigate the validity of the contents of the present paper for the case of spaces over \textit{arbitrary} non-archimedean and non-trivially valued complete fields.