Strichartz estimates on the quaternion Heisenberg group (Q2437725)

The purpose of the present manuscript focuses on showing dispersive and Strichartz estimates associated to the following Cauchy problem for the wave equation on \(\mathbb {H}\times \mathbb R^3\) (the quaternion Heisenberg group, where \(\mathbb H\) is the 4-dimension quaternion algebra): \[ (*)_f:\quad \left(\frac{\partial}{\partial t}-\Delta_{\mathbb {H}\times \mathbb R^3}\right)u = f,\quad \text{and}\quad u_{\big|_{t=0}}=u_0,\;\frac{\partial u}{\partial t}_{\big|_{t=0}}=u_1, \] where \(\Delta_{\mathbb {H}\times \mathbb R^3}\) stands for the Kohn-Laplacian operator defined explicitly. In the second section, the authors recall some properties of \(\mathbb {H}\times \mathbb R^3\), such as the formulæ\ of the left-invariant vector field generators of its associated Lie algebra, its unitary representation on the Fock space, and the (inverse) Fourier transform of tempered distributions. The third section focuses on the expression of the Littlewood-Paley inequality (including its inverse) on \(\mathbb {H}\times \mathbb R^3\) (Theorem 3.4). As regards the fourth section, the authors recall some facts on a family of homogeneous Besov spaces and their related norms. Then they show some Sobolev embedding theorems on \(\mathbb {H}\times \mathbb R^3\) (Theorems 4.9 and 4.10). Also, they prove when a Besov space coincides with the Sobolev space of finite order (Theorem 4.12). The fifth section contains the proofs of the two main results. The first one is on the dispersive estimates. The authors state that the \(L^{\infty}(\mathbb {H}\times \mathbb R^3)\)-norm of the solution of \((*)_0\) is bounded, up to a multiplicative constant, by a decreasing function of \(t\) times the sum of the Besov-norm of \(u_0\) and \(u_1\), respectively (Theorem 5.1). The proof is based on a series of lemmas. In the second result, the authors show the Strichartz estimates by considering the conjugate Hölder numbers \(4p\) and \(\frac{2q}{5}\). For a positive number \(T\), they show that the \(L_p([0,T])\otimes L_q(\mathbb {H}\times \mathbb R^3)\)-norm of a solution of \((*)_f\) is bounded by the \(L_1([0,T])\otimes L_2(\mathbb {H}\times \mathbb R^3)\)-norm plus an explicit positive function of \(u_0, u_1\) (Theorem 5.9). The proof is based on writing a solution of \((*)_f\) as the sum of a solution of \((*)_0\) endowed with nonzero initial data and a solution of \((*)_0\) such that \(u_0=u_1=0\).