Isometries of Minkowski geometries (Q332644): Difference between revisions
From MaRDI portal
Revision as of 21:17, 12 July 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Isometries of Minkowski geometries |
scientific article |
Statements
Isometries of Minkowski geometries (English)
0 references
8 November 2016
0 references
Let \(V\) be a finite dimensional real Banach space. There exists a semi-inner-product (namely, a function \([\cdot,\cdot]: V\times V\to \mathbb{R}\) which is linear in its first variable, strictly positive and satisfies the Schwartz inequality) satisfying \(\|x\|=[x,x]\). A linear operator \(A\) on \(V\) is called adjoint abelian if it satisfies \([Ax, y] = [x, Ay]\) for every \(x, y\in V\). Theorem 5 in this paper shows that, if \(V\) is a smooth and \(A\) is adjoint abelian, then \(V\) can be decomposed into the direct sum of \(A\)-invariant subspaces of dimension at most two, and the restriction of \(A\) to a 2-dimensional component is a generalized dilation. Besides, if there exists \(k\in \mathbb{R}\) such that \(|[x,y]-[x,z]|\leq k \|y-z\|\) when \(x,y,z\) belong to the unit ball, it is conjectured that \(A\) is diagonalizable (the case \(V=l_p\), \(1<p<\infty\), is proved). Theorem 10 proves that, if \(U\) is an isometry on \(V\) that preserves the semi-inner-product, then \(V\) can be decomposed into the direct sum of \(U\)-invariant mutually orthogonal 1-dimensional subspaces with eigenvalue \(\pm 1\), and an even number of \(U\)-invariant 2-dimensional subspaces. The \(2\)-dimensional subspaces are orthogonal to the 1-dimensional ones, and the restriction of \(U\) to a 2-dimensional component is a generalized rotation. In the last section of the paper, it is proved that, if the unit ball of a 3-dimensional Minkowski space \(V\) does not intersect a 2-plane in an ellipse, then the group of isometries of \(V\) is isomorphic to the semi-direct product of the translation of \(\mathbb{R}^3\) with a finite group \(G\) of linear transformations with determinant \(\pm 1\). The author asks if \(G\) can be replaced by the special linear group when the dimension is greater than 2 and \(V\) has no 2-dimensional Euclidean subspace.
0 references
adjoint abelian operator
0 references
Banach space
0 references
Minkowski geometry
0 references
semi-inner product
0 references
isometry group
0 references