Description of the duals of subspaces of infinitely differentiable functions (Q6564077)

Let \(D\) be a closed bounded convex domain in some \(\mathbb R^{p}\) and \(h\) a convex function on \(D\) with \(h(0)=0\). Denote by \(C(D,\mathcal{M})\) the space of infinitely differentiable functions on \(D\) which extend with all their derivatives to the closure of \(D\), equipped with the norm,\N\[\N\left\Vert f\right\Vert _{C(D,\mathcal{M})}=\sup_{n\in \mathbb Z_{+}^{p}}\frac{ \sup_{x\in \overline{D}} \left\vert D^{n}\left( f\right) \left( x\right) \right\vert }{M_{n}}\N\]\Nwhere \(\mathcal{M=}\left( M_{n}\right) _{n\in \mathbb Z_{+}^{p}}\), and \( M_{n}\circeq e^{h(n)}\), \(n\in \mathbb Z_{+}^{p}\).\N\NNext, consider a family \(\Sigma \) of sequences \(\mathcal{M}^{m}\circeq \left( M_{n}^{m}\right) _{n\in \mathbb Z_{+}^{p}},m\in \mathbb N\), with some additional conditions on \(\left( M_{n}^{m}\right) _{n\in \mathbb Z_{+}^{p}, m\in \mathbb N}\) ensuring, in particular, compact continuous inclusions of \(C(D,\mathcal{M}^{m})\) into \(C(D,\mathcal{M}^{m+1})\) for all \(m\), and the containment of the function class \(\mathrm{EXP} =\left( e^{\langle \ast ,\, z\rangle }\right) _{z\in {C}^{p}}\), \(\langle x, z\rangle \circeq \sum_{i=1}^{p}x_{i}z_{i}\), in each \(C(D,\mathcal{M}^{m})\), \( m\in \mathbb N\). Form \(C(D,\Sigma )\circeq \bigcup _{m} C(D,\mathcal{M}^{m})\) and equip it with the natural inductive limit topology.\N\NThis article aims to describe the image of the Fourier-Laplace transformation, \(S\rightarrow \widehat{S}(z)\circeq S(e^{\langle \ast ,\,z\rangle })\), defined on the dual of \(C(D,\Sigma )\) with strong dual topology as the topological projective limit \(H(\Phi)\), \(\Phi =\left( \rho _{n}\right) _{n}\), \(\rho _{n}:C^{p}\rightarrow \mathbb R^{+}\), \(n\in \mathbb N\), of weighted spaces of entire functions:\N\[\N\left( P_{n},^{\left\Vert \ast \right\Vert _{n}}\right) =\left( f\in H\left( C^{p}\right) :\left\Vert f\right\Vert _{n}\circeq \sup_{z\in C^{p}}\frac{\left\vert f\left( z\right) \right\vert }{\rho _{n}\left( z\right) }\right) ,\quad n\in \mathbb N.\N\]\N\NThe main theorem of the paper states, under certain conditions on the \(h_{n}\), that the Fourier-Laplace transformation is a topological isomorphism from \( C(D,\Sigma )\), \(\Sigma =\left\{ \mathcal{M}^{m}\right\} _{m}\), onto the weighted space of entire functions H(\(\Phi \)), \(\Phi =\left( \varphi _{m}\right) _{m}\),\N\[\N\varphi _{m}\left( z\right) \circeq H_{D}\left( \textrm{Re }z\right) +\ln A_{m}\left( z\right) ,\ z\in C^{p},\ m\in \mathbb N\N\]\Nwhere \(H_{D}\) is the support function of \(D\) and \(A_{m}\left( z\right) \circeq \sum_{k}\) \(\frac{\left\vert z^{k}\right\vert }{M_{k}^{m}}\), \( z\in C^{p}\).\N\NThe proof uses \textit{B. A. Taylor} [Commun. Pure Appl. Math. 24, 39--51 (1971; Zbl 0205.41501)] and \textit{M. Neymark}'s [Ark. Mat. 7, 577--594 (1969; Zbl 0172.42101)] scheme and some previous results of \textit{I.~K. Musin} and \textit{P.~V. Yakovleva} [Cent. Eur. J. Math. 10, No.~2, 665--692 (2012; Zbl 1252.46025)].