Inverse function in the neighborhood of an abnormal point of a smooth map (Q1761007)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Inverse function in the neighborhood of an abnormal point of a smooth map
scientific article

    Statements

    Inverse function in the neighborhood of an abnormal point of a smooth map (English)
    0 references
    15 November 2012
    0 references
    Suppose that \(X=\mathbb{R}^n\) and \(Y=\mathbb{R}^k\) are arithmetic spaces and we have given a~smooth map \(F:X\rightarrow Y\) and a point \(x_0\in X\). In the paper there are presented conditions on \(F\) under which a neighborhood of point \(y_0=F(x_0)\) exists such that for any \(y\) from this neighborhood the equation \(F(x)=y\) has a solution \(x=\varphi(y)\) for which \(\varphi(y_0)=x_0\) and the map \(\varphi\) is continuous at point \(y_0\). Two types of points \(x_0\) are discussed in this problem: -- normal points for which \(\mathrm{im}\,F'(x_0)=Y\) (im denotes here the image of linear operator), -- abnormal points for which \(\mathrm{im}\,F'(x_0)\neq Y\). If the point \(x_0\) is normal, the respective conditions are given in the classical inverse function theorem. If the point \(x_0\) is abnormal, then this theorem does not hold. Therefore, the new type of sufficient conditions are needed here. The authors: Avakov, Arutyunov and Karamzin are specialists in this subject and the presented paper represents the summary of their previous research. In the paper six theorems and two essential examples are presented. The first theorem concerns a weakened type of normal points \(x_0\), when there exists a direction \(h\) along which the map \(F\) is 2-regular at point \(x_0\). The second theorem is connected with abnormal points for which \(\mathrm{codim\,im}\,F'(x_0)=1\). The other theorems concern the abnormal points with \(\mathrm{codim\,im}\,F'(x_0)\geq 2\). The proofs of these theorems are based on a theory of necessary optimality conditions. An important feature distinguishing these theorems from the first and the second one is that the function \(x(\cdot)\), satisfying equation \(F(x(y))=y\), cannot be chosen any more such that it is continuous in any neighborhood of \(x_0.\) Moreover, the map \(F\) under the assumptions of these theorems may not have any directions in which it is 2-regular at \(x_0\). Respective examples are also given and discussed. In my opinion, the paper is worth to look over and the presented results have a chance to be a permanent achievement in the subject.
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references