MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I am a mathematical physicist and I am studying certain discrete dynamical systems defined in terms of piecewise linear mappings, which may be expressed in terms of expressions over the max-plus semi-field.

In the study of the points of non-differentiability of these piecewise linear mappings, I was wondering, is there an established proof that (in n-dimensions) every set of points of non-differentiability of a piecewise linear mapping (with rational slopes) may be expressed as the image of a variety over a non-archimedean valuation field under the endowed valuation?

Is there some, possibly constructive proof, with some sort of standard construction?

share|cite|improve this question
Could you name your spaces, so I can understand what you are asking? Right now, it looks like you start with a PL map from $\mathbb{R}^n$ to itself, and are considering the subspace $X$ of the domain on which the map is non-differentiable. Am I correct that you want to know if $X$ is the image under the valuation map of some variety in $\mathbb{A}^n_K$, where $K$ is a field endowed with a nonarchimedean real valuation? – S. Carnahan Oct 19 '10 at 3:35
Indeed, you are correct. I have $X \subset \mathbb{R}^n$ which is the set of points in which a PL mapping, $f : \mathbb{R}^n \to \mathbb{R}^m$, is not differentiable. One may replace $\mathbb{R}$ with $S = (\mathbb{R} \cup \{\infty\}, \max, + )$. Now, the field that seems popular would be letting $\mathbb{K}$ be the set of algebraic functions (with usual index map). Not being an expert in the field, I would like to know if it has been established that $X$ is the image variety over $\mathbb{A}_{\mathbb{K}}^n$. Is there, for example, a standard construction of an underlying ideal? – Chris Ormerod Oct 22 '10 at 1:04

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.