The Dolbeault complex in infinite dimensions II

From MaRDI portal





A well-known theorem asserts that for any continuous \(\overline \partial\)-closed (0,1)-form \(f\) on the unit ball of a finite dimensional complex normed space \(E\) there exists a \(C^1\) function \(u\) on that ball satisfying \(\overline \partial u=f\). Since the construction by G. Coeuré of a \(C^1\) \(\overline \partial\)-closed (0,1)-form on \(l^2\) for which there is no fonction \(u\) solving \(\overline \partial u=f\) on any nonempty open set, the question was: for what kind of normed spaces and under what regularity assumptions does this theorem remain valid? This paper presents the first (and only) positive result in this direction. Namely, such a function \(u\) does exist when \(E=l^1\) and \(f\) is Lipschitz continuous on all balls of radius \(r<1\). The proof goes in two steps involving very interesting tools. In the first step a solution \(u\) is defined on any ball of radius \(r<1\). This uses a monomial expansion \(f\sim \sum f_k\) \((k\in \mathbb{Z}^\mathbb{N})\) and a Cesàro-Féjer summation process for suitable \(u_k\) solving \(\overline\partial u_k= f_k\). The second step classically uses an approximation result to construct the solution \(u\) on the whole ball. Precisely it is proven that any holomorphic function on the unit ball of \(l^1\) can be approximated in some sense by entire functions. The end of the paper is devoted to a generalization of Coeuré's example giving a \(C^{p-1}\) \(\overline \partial\)-closed (0,1)-form \(f\) on \(l^p\) \((p=1,2, \dots)\) for which there is no function \(u\) solving \(\overline u=f\) on any nonempty open set. [For part I see ibid. 11, No. 3, 485-520 (1998; Zbl 0904.32014)].




Cited in
(30)








This page was built for publication: The Dolbeault complex in infinite dimensions II

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4243742)