0
$\begingroup$

Let the zero sets $F=\{x \in \mathbb{R}^n: f(x) = 0\}$, $G = \{x \in \mathbb{R}^n : g(x) = 0\}$, where $f$ and $g$ are $m$-dimensional real, analytic, continuous, and nonlinear vector functions. Under some assumptions, these sets define hypersurfaces of zero Lebesgue measure in $\mathbb{R}^n$. I was wondering:

  1. Are $F$ and $G$ always submanifolds embedded in $\mathbb{R}^n$ or are there exceptions - in the latter case, are there conditions that guarantee that they are submanifolds ?
  2. What is the dimension of $F\cap G$ ? As pointed out here Measure of the intersection of two manifolds, if $F$ and $G$ are $(n-1)$-dimensional manifolds and their intersection is transversal, then $\text{dim}(F\cap G) = n-2$. However, is there anything that can be said if the intersection is not transversal ? In general, I am interested in some sort of inequality $\text{dim}(F\cap G) \leq n-2$, assuming that $F \subseteq G$, that $G \subseteq F$ do not hold.
$\endgroup$
4
  • $\begingroup$ I think you should edit your question. What are your assumptions about $f$ and $g$? And by measure: Do you mean Hausdorff measure? Lebesgue measure? $\endgroup$ Oct 14, 2020 at 10:24
  • $\begingroup$ Thanks! I meant Lebesgue measure. Regarding the functions, besides continuity, they are real and analytic. What other assumptions would make answering the question easier ? $\endgroup$
    – Chris
    Oct 14, 2020 at 10:50
  • $\begingroup$ You'd need $f$ and $g$ of maximal rank (eg submersions). Note that $f$ and $g$ define a function $(f,g)\colon \mathbb{R}^n \to \mathbb{R}^{2m}$ and $F\cap G$ is the zero-set of this function. If $F$ and $G$ are $(n-1)$-dimensional, then $m=1$. So I would imagine that away from singularities, $F\cap G$ is $(n-2)$-dimensional. $\endgroup$
    – David Roberts
    Oct 14, 2020 at 11:40
  • $\begingroup$ Thanks! I didn't really get why $m=1$ but then I guess $F$ and $G$ are equivalent to the sets $\{x:\|f(x)\| = 0\}$, $\{x:\|g(x)\| = 0\}$, which are zero-sets of real 1D functions. If $f$, $g$ are close to singularities, what would that imply for $F\cap G$ ? In general, is there any case when the dimension of $F\cap G$ would be $n-1$ ? $\endgroup$
    – Chris
    Oct 14, 2020 at 12:20

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.