Symmetric preparation of systems
Nils Dencker
TL;DR
This work extends the Weierstrass and Malgrange preparation theorems to symmetric matrix-valued systems. For a smooth or analytic symmetric $N\times N$ matrix function $F(t,x)$ vanishing to first order with $\partial_t F(0,0) > 0$ and $F(0,0)=0$, it establishes the existence of a local symmetric factorization $F(t,x)=U(t,x)\bigl(t\operatorname{Id}_N+M(x)\bigr)U^*(t,x)$ with $|U(0,0)|\neq0$ and $M(x)$ symmetric, $M(0)=0$, and proves uniqueness in the symmetric case ($U$ positive definite). When $F(0,0)\neq0$, the remainder $F(0,0)$ can be accommodated, yielding a version with remainder. The analytic case is proved via an analytic inverse function theorem, while the smooth case uses the Nash–Moser theorem, aided by a symmetric Malgrange division, to obtain locally solvable and tame mappings. Together, these results generalize scalar preparation to symmetric matrix systems, providing a canonical, symmetry-preserving factorization with implications for spectral theory and operator analysis.
Abstract
In this paper we generalize the Weierstrass and Malgrange preparation theorems to the symmetric matrix valued case, proving symmetric preparation of analytic and smooth symmetric systems that vanish of first order.
