On the bifurcation set of a polynomial function and Newton boundary (Q1173974): Difference between revisions

It is known that for a polynomial function \(f: {\mathbb{C}}^ n\rightarrow{\mathbb{C}}\) there exists a finite set \(\Gamma\subseteq{\mathbb{C}}\), such that \(f: {\mathbb{C}}^ n\backslash f^{- 1}(\Gamma)\rightarrow{\mathbb{C}}\backslash\Gamma\) is a locally trivial fibration. The smallest such set \(\Gamma\) is called bifurcation set, denoted by \(B_ f\). If the map \(f\) is not proper, the set \(B_ f\) contains besides the set \(\Sigma_ f\) of all critical values of \(f\) perhaps some other points. Let \({\mathcal M}(f)\) be the Milnor set \({\mathcal M}(f)=\{z\in{\mathbb{C}}^ n:\hbox{there exists } \lambda\in \mathbb{C}\hbox{ such that grad} f(z)=\lambda z\}\) and let \(S_ f\) be defined by: \(S_ f=\{c\in \mathbb{C}:\hbox{there exists a sequence }\{z^ k\}_ k\subseteq M(f)\) such that \(\lim_{k\rightarrow\infty}\| z^ k\|=\infty\) and \(\lim_{k\rightarrow\infty} f(z^ k)=c\}\). Then the following theorems are valid: Theorem 1. Let \(f: {\mathbb{C}}^ n\rightarrow{\mathbb{C}}\) be a polynomial map. Then \(B_ f\subseteq\Sigma_ f\cup S_ f\). Theorem 2. If \(f\) is not convenient, Newton nondegenerate and \(f(0)=0\), then \(B_ f\subseteq\Sigma_ f\cup\{0\}\cup\bigcup_{\Delta\in{\mathcal B}}\Sigma_ \Delta\), where \(\mathcal B\) is the set of bad faces of \(\overline{\hbox{supp}(f)}\) and \(\Sigma_ \Delta=\{f_ \Delta(z^ 0):\;z^ 0\in(\mathbb{C}^*)^ n\) and \(\hbox{grad} f_ \Delta(z^ 0)=0\}\).