Abstract:We discussed some properties of a family of symmetric spaces, namely $\mathcal{P}(n):=SL(n,\mathbb{R})/SO(n,\mathbb{R})$, where we replace the Riemannian metric on $\mathcal{P}(n)$ with a premetric suggested by Selberg. These properties are generalizations of properties on the hyperbolic spaces $\mathbf{H}^n$ related to Poincaré's fundamental polyhedron theorem.
What problem does this paper attempt to address?
This paper attempts to solve a series of problems related to the symmetric space \(P(n)=\text{SL}(n,\mathbb{R})/\text{SO}(n,\mathbb{R})\). In particular, when the pre - metric proposed by Selberg is used on \(P(n)\) to replace the original Riemannian metric, some properties of these symmetric spaces are studied. Specifically, the paper focuses on the following aspects:
1. **Intersection Criteria of Selberg Bisectors**: Study the intersection conditions between Selberg bisectors, including how to determine whether two bisectors intersect. This involves the properties of matrix pencils and the analysis of their generalized eigenvalues.
2. **Shape of Dirichlet - Selberg Domains**: Explore the geometric shape of Dirichlet - Selberg domains defined on \(P(n)\), especially the properties under the Selberg pre - metric. This includes the classification of Dirichlet - Selberg domains with finite and infinite edges.
3. **Angle Function**: Propose an angle function defined on \(P(n)\) to describe the "angle" between Selberg bisectors. This angle function satisfies specific natural properties, such as invariance, symmetry, and the triangle inequality.
4. **Algorithm Implementation**: Develop an algorithm for calculating the finite - edged polyhedra of the partially ordered set (poset) structure, which is an important part of the Poincaré algorithm. The application of this algorithm on \(P(n)\) is similar to the Epstein and Petronio algorithm in hyperbolic space.
5. **Existence of Schottky Subgroups**: Prove the existence of Schottky subgroups in \(\text{SL}(n,\mathbb{R})\), and the Dirichlet - Selberg domains of these subgroups have pairwise disjoint faces, similar to the Schottky groups in \(\text{SO}^+(n,1)\).
### Main Results
- **Theorem 3.1**: Propose an invariant angle function to describe the "angle" between Selberg bisectors. According to different cases of the generalized eigenvalues of matrix pencils, three different expressions are given:
- When the generalized eigenvalues contain non - real numbers, the angle function is:
\[
\theta(\sigma_1,\sigma_2)=\frac{1}{k}\sum_{i = 1}^k|\arg(\lambda_i)|
\]
- When all different generalized eigenvalues are real numbers or infinity, the angle function is:
\[
\theta(\sigma_1,\sigma_2)=\arccos\left(\frac{\sum_{i = 1}^k\frac{\lambda_{i + 1}+\lambda_i}{\lambda_{i + 1}-\lambda_i}}{\sqrt{\left(\sum_{i = 1}^k\frac{1}{\lambda_{i + 1}-\lambda_i}\right)\left(\sum_{i = 1}^k\frac{(\lambda_{i + 1}+\lambda_i)^2}{\lambda_{i + 1}-\lambda_i}\right)}}\right)
\]
- When the matrix pencil has at most two different generalized eigenvalues and they are all real numbers, there is no invariant angle function.
- **Theorem 4.1**: Provide criteria for determining whether two hyperplanes intersect. Specifically, two hyperplanes \(A^{\perp}\) and \(B^{\perp}\) do not intersect if and only if one of the following conditions holds:
- The matrix pencil \((A,B)\) is diagonal and semi - positive