Searcharxiv⌕ Search

arXiv subjects

Krzysztof Kurdyka

Publications and source records attributed to Krzysztof Kurdyka.

20 records · Page 2Linked to original sources

Hyperbolic polynomials and multiparameter real analytic perturbation theory

Let $P(x,z)= z^d +\sum_{i=1}^{d}a_i(x)z^{d-i}$ be a polynomial, where $a_i$ are real analytic functions in an open subset $U$ of $\R^n$. If for any $x \in U$ the polynomial $z\mapsto P(x,z)$ has only real roots, then we can write those roots as locally lipschitz functions of $x$. Moreover, there exists a modification (a locally finite composition of blowing-ups with smooth centers) $σ: W \to U$ such that the roots of the corresponding polynomial $\tilde P(w,z) =P(σ(w),z), w\in W $, can be written locally as analytic functions of $w$. Let $A(x), x\in U$ be an analytic family of symmetric matrices, where $U$ is open in $\R^n$. Then there exists a modification $σ: W \to U$, such the corresponding family $\tilde A(w) =A(σ(w))$ can be locally diagonalized analytically (i.e. we can choose locally eigenvectors in an analytic way). This generalizes the Rellich's well known theorem (1937) for one-parameter families. Similarly for an analytic family $A(x), x\in U$ of antisymmetric matrices there exits a modification $σ$ such that we can find locally a basis of proper subspaces in an analytic way.

math.GM↗

Proof of the gradient conjecture of R. Thom

Let x(t) be a trajectory of the gradient of a real analytic function and suppose that x_0 is a limit point of x(t). We prove the gradient conjecture of R. Thom which states that the secants of x(t) at x_0 have a limit. Actually we show a stronger statement: the radial projection of x(t) from x_0 onto the unit sphere has finite length.

math.AG↗