Characterization of the simple \(L^1(G)\)-modules for exponential Lie groups. (Q1880102): Difference between revisions

Let \(G = \exp \mathfrak g\) be a connected, simply connected exponential solvable Lie group and \(\widehat G\) be the unitary dual of \(G\). So, it is well known [\textit{H. Leptin} and \textit{J. Ludwig}, Unitary representation theory of exponential Lie groups (1994; Zbl 0833.22012)] that every element \(\pi \in \widehat G\) is obtained up to equivalence by considering the representation \(\text{Ind}_P^G\chi_l\) where \(l\in \mathfrak g^*\), the dual space of \(\mathfrak g\), and \(P\) is a Pukanszky polarization at \(l\) and \(\chi_l\) the unitary character of \(P\). When \(G\) is nilpotent, then every simple \(L^1(G)\) module is equivalent to a module of the form \((\pi|_{\mathcal{H}_\pi^0}, \mathcal{H}_\pi^0)\), where \(\pi \in \widehat G,\;\mathcal{H}_\pi\) is the Hilbert space of \(\pi\) and \[ \mathcal{H}_\pi^0 = \text{span}-\{\pi(f)\xi,\;f\in L^1(G),\;\pi(f) \;\text{of finite rank} ,\;\xi\in \mathcal{H}_\pi\}. \] The aim of the paper under review is to generalize such characterization for arbitrary exponential solvable Lie groups. The authors define for every \( \overline{p}=(p_1,\ldots, p_m)\in [1, +\infty]^m,\;l\in \mathfrak g^*\) and \(P\) a Pukanszky polarization at \(l\) a topologically irreducible representation \((\pi_{l, P, \overline{p}}, \mathcal{H}_{l, P, \overline{p}})\) by induction on the \(L^{\overline{p}}\)-space using a well chosen norm \(\| \cdot\| \) on this space. They show that every representation \(\pi_{l, P, \overline{p}}\) may be characterized up to equivalence by the \(G\)-orbit of a couple \((l, \nu)\) where \(l\in \mathfrak g^*\) and \(\nu\) is a real linear form on \(\mathfrak g(l) / \mathfrak g(l)\cap \mathfrak n\) where \(\mathfrak g(l)\) is the radical of the bilinear form \(B_l(X, Y) = l([X, Y]) \) and \(\mathfrak n\) is the nilradical of \(\mathfrak g\). They obtain the following description: every simple \(L^1(G)\)-module is equivalent to a module of the form \((\pi_{l, P, \overline{p}}, \mathcal{H}_{l, P, \overline{p}}^0)\) where as above \[ \mathcal{H}_{l, P, \overline{p}}^0 = \text{span}-\{\pi_{l, P, \overline{p}}(f)\xi,\;f\in L^1(G),\;\pi_{l, P, \overline{p}}(f) \;\text{of finite rank}\;, \xi\in \mathcal{H}_{l, P, \overline{p}}\}. \]