Sharp \(L^2\)-Caffarelli-Kohn-Nirenberg inequalities for Grushin vector fields (Q2199967)

On \(\mathbb{R}^{N+1}\) with coordinates \(z \in \mathbb{R}^{N}\) and \(t\in \mathbb{R}\) define the Grushin gradient \(\nabla_G=(\nabla_z,|z|\partial_t)\) and the Grushin operator \(\Delta_G= \Delta_z+|z|^2\partial_t^2\) associated to the vector fields \(X_j = \partial_{z_j}\), \(j=1, \dots, N\), and \(T=|z|\partial_t\). In this paper are characterized the best constants and minimizers for inequalities \[C \int_{\mathbb{R}^{N+1}} |u|^2 \frac{\psi}{\rho^{a+b+1}}\ dz\,dt \le \left( \int_{\mathbb{R}^{N+1}} \frac{|\nabla_G u|^2 }{\rho^{2b}}\ dz\,dt \right)^\frac12 \left( \int_{\mathbb{R}^{N+1}} |u|^2 \frac{\psi}{\rho^{2a}}\ dz\,dt \right)^\frac12, \] where \((a,b) \in \mathbb{R}^2\), \(\psi = \frac{|z|^2}{\rho^2}\), and \(\rho= \left( |z|^{4}+4t^{2}\right)\). This analysis is done by a clever use of special polar coordinates and by taking advantage of the commutation relations between \(\Delta_{G}\) and an anisotropic Kelvin transform.