The Cayley Transform on Representations

Jingyu Lu,Ke Ye
2024-11-04
Abstract:The classical Cayley transform is a birational map between a quadratic matrix group and its Lie algebra, which was first discovered by Cayley in 1846. Because of its essential role in both pure and applied mathematics, the classical Cayley transform has been generalized from various perspectives. This paper is concerned with a representation theoretic generalization of the classical Cayley transform. The idea underlying this work is that the applicability of the classical Cayley transform heavily depends on how the Lie group is represented. The goal is to characterize irreducible representations of a Lie group, to which the classical Cayley transform is applicable. To this end, we first establish criteria of the applicability for a general Lie group. If the group is semisimple, we further obtain a geometric condition on the weight diagram of such representations. Lastly, we provide a complete classification for classical complex simple Lie groups and their compact real forms. Except for the previously known examples, spin representations of $\mathrm{Spin}(8)$ are the only ones on our list.
Representation Theory
What problem does this paper attempt to address?
The problem that this paper attempts to solve is about the generalization of the Cayley transform in representation theory and the conditions for its applicability. Specifically, the author focuses on how to characterize the irreducible representations of Lie groups through the Cayley transform, especially those representations to which the classical Cayley transform can be applied. The core problem of the paper is: **Problem 1.2 (Lemire - Popov - Reichstein problem in representation theory)**: For a Lie group \(G\) and its representation \(\rho: G \to GL(V)\), which pairs \((G, \rho)\) can make the classical Cayley transform \(C\) applicable? Here, the applicability means that there exists an open neighborhood \(V\) of \(\mathfrak{g}\) such that \(C\circ d\rho(V)\subseteq\rho(G)\). ### Main Results The main contribution of the paper lies in establishing several equivalent conditions for the applicability of the Cayley transform and classifying specific types of Lie groups. Specifically: 1. **Equivalence between Applicability and the Power - Generation Property**: - Theorem 4.9 establishes several equivalent conditions for \(\rho\) to be a Cayley representation, including that \(d\rho(\mathfrak{g})\) has the power - generation property. - The power - generation property means that every element \(u\) in \(\mathfrak{u}\) satisfies \(u^{2k + 1}\in\mathfrak{u}\) for any integer \(k\geq0\). 2. **The Case of Semisimple Lie Groups**: - Theorem 5.2 further discusses the case of the semisimple Lie group \(G\), where \(\mathfrak{g}_0=\mathfrak{h}\) and \(\mathfrak{g}=\text{Ad}(G)(\mathfrak{h})\). In this case, \(\rho\) is a Cayley representation if and only if \(\tilde{d\rho}(S^3\mathfrak{h})\subseteq d\rho(\mathfrak{h})\). 3. **Classification of Irreducible Representations**: - Section 6 provides a complete classification of the irreducible representations of classical complex simple Lie groups and their compact real forms, and determines which representations can apply the Cayley transform. The main results are as follows: - \(G\) is of type \(A_1, B_n (n\geq1), C_n (n\geq3)\) or \(D_n (n\geq3)\), and \(\rho\) is the standard representation. - \(G\) is of type \(D_4\), and \(\rho\) is one of the two spin representations. ### Mathematical Formulas - **Cayley Transform**: \[ C: \mathfrak{gl}(V)\to GL(V),\quad C(u)=(I_n + u)(I_n - u)^{-1} \] - **Power - Generation Property**: \[ u^{2k+1}\in\mathfrak{u}\quad\text{for any integer }k\geq0 \] - **Condition for Irreducible Representations**: \[ \tilde{d\rho}(S^3\mathfrak{h})\subseteq d\rho(\mathfrak{h}) \] These results not only provide a profound theoretical understanding of the applicability of the Cayley transform but also have important significance in practical applications, especially in numerical linear algebra, differential equation solving, and optimization.