Optimal Sobolev inequalities in the hyperbolic space

Zdeněk Mihula
2023-05-11
Abstract:We find the optimal function norm on the left-hand side of the $m$th order Sobolev type inequality $\|u\|_{Y(\mathbb{H}^n)} \leq C \|\nabla_g^m u\|_{X(\mathbb{H}^n)}$ in the $n$-dimensional hyperbolic space $\mathbb{H}^n$, $1\leq m < n$. The optimal function norm in the inequality among all rearrangement-invariant function norms is completely characterized. A variety of concrete examples of optimal function norms in the inequality is provided. The examples include delicate limiting cases and, especially when $m\geq3$, seem to provide new, improved inequalities in these limiting cases.
Functional Analysis
What problem does this paper attempt to address?
### Problems the paper attempts to solve This paper aims to study the high - order Sobolev - type inequalities in the \(n\)-dimensional hyperbolic space \(H^n\) and find the optimal form of the function norms on the left - hand side of these inequalities. Specifically, the paper focuses on Sobolev - type inequalities in the following form: \[ \|u\|_{Y(H^n)} \leq C \| \nabla^m_g u \|_{X(H^n)} \] where: - \(1 \leq m < n\) - \( \| \cdot \|_{Y(H^n)} \) and \( \| \cdot \|_{X(H^n)} \) are rearrangement - invariant function norms (such as Lebesgue norms, Lorentz norms or Orlicz norms) - \(V^m_0X(H^n)\) is a Sobolev - type function space, which contains functions that vanish at infinity together with their lower - order derivatives The main objectives of the paper are: 1. **Determine the optimal function norms**: Given \(X(H^n)\), find the optimal (i.e., the smallest) rearrangement - invariant function space \(Y(H^n)\) such that the above inequality holds. 2. **Provide specific examples**: In particular, in the case of \(m\geq3\), provide some new and improved inequalities. ### Main results The main results of the paper include: - **General theorem**: In the most general case, the paper fully characterizes the optimal rearrangement - invariant function norms without imposing unnecessary assumptions on \(X(H^n)\). - **Simplified description**: When \(X(H^n)\) is not "close" to \(L^1(H^n)\) or \(L^\infty(H^n)\), the description of the optimal function space can be greatly simplified. - **Specific examples**: Many specific examples of optimal rearrangement - invariant function norms are provided, especially when \(X(H^n)\) belongs to the class of Lorentz - Zygmund spaces. ### Specific examples 1. **For \(u\in V^m_0L^1(H^n)\)**: - When \(m\geq3\) and \(m\) is odd: \[ \int_0^1 t^{n - m}u^*(t)\,dt+\sup_{t\in[1,\infty)}tu^{**}(t)(1 + \log t)^{\frac{m - 1}{2}}\leq C \| \nabla^m_g u \|_{L^1(H^n)} \] - When \(m\) is even: \[ \sup_{t\in(0,1)}t^{n - m}u^{**}(t)+\sup_{t\in[1,\infty)}tu^{**}(t)(1 + \log t)^{\frac{m}{2}}\leq C \| \nabla^m_g u \|_{L^1(H^n)} \] 2. **For \(u\in V^m_0L^{\frac{n}{m}}(H^n)\)**: \[ \left(\int_0^1\left(\frac{u^{**}(t)}{1-\log t}\right)^{\frac{n}{m}}\frac{dt}{t}+\int_1^\infty(u^{**}(t))^{\frac{n}{m}}\,dt\right)^{\frac{m}{n}}\leq C \| \nabla^m_g u \|_{L^{\frac{n}{m}}(H^n)} \] 3. **When \(X(H^n)=L^\infty(H^n)\)**: \[ \| u \|