| Popis: |
For a regular (in a sense) mapping $v:\mathbb{R}^n \to \mathbb{R}^d$ we study the following problem: {\sl let $S$ be a subset of $m$-critical a set $\tilde Z_{v,m}=\{{\rm rank} \nabla v\le m\}$ and the equality $\mathcal{H}^\tau(S)=0$ (or the inequality $\mathcal{H}^\tau(S)<\infty$) holds for some $\tau>0$. Does it imply that $\mathcal{H}^{\sigma}(v(S))=0$ for some $\sigma=\sigma(\tau,m)$?} (Here $\mathcal{H}^\tau$ means the $\tau$-dimensional Hausdorff measure.) For the classical classes $C^k$-smooth and $C^{k+\alpha}$-Holder mappings this problem was solved in the papers by Bates and Moreira. We solve the problem for Sobolev $W^k_p$ and fractional Sobolev $W^{k+\alpha}_p$ classes as well. Note that we study the Sobolev case under minimal integrability assumptions $p=\max(1,n/k)$, i.e., it guarantees in general only {\it the continuity} (not everywhere differentiability) of a mapping. In particular, there is an interesting and unexpected analytical phenomena here: if $\tau=n$ (i.e., in the case of Morse--Sard theorem), then the value $\sigma(\tau)$ is the same for the Sobolev $W^k_p$ and for the classical $C^k$-smooth case. But if $\tauComment: arXiv admin note: text overlap with arXiv:1706.05266 |
