scientific article; zbMATH DE number 3880012

From MaRDI portal
Publication:3345693





The definitions of topological (or mapping) degree and its calculation are synthetically described in this paper. Section 2 describes the combinatorial definition of the mapping degree \(\deg\phi\) and points out that \(\deg\phi\) only depends on the mapping \(\phi: | K| \to | L|\), and does not depend on how to subdivide the spaces \(| K|\) and \(| L|\) into \(K\) and \(L\). Section 3 describes the analytical definition of the mapping degree \(\deg f\) and points out that \(\deg f\) indicates the situation of the solution of the equation \(fx=y\) where \(y\) is a regular value, that is, a nonzero \(\deg f\) shows that the equation \(fx=y\) has solutions, where \(y\) is a regular value. In section 4 an algorithm for the calculation of \(\deg(F,P,0)\) where \(F=(f_ 1,\ldots,f_ n): P\to\mathbb R^ n,\) differentiable or continuous, and \(P\) is the unit spheroid in \(\mathbb R^ n\). \(\deg(F,P,0)\), the mapping degree of \(F\) at zero on \(P\), indicates the situation of zero points of \(F\) inside \(P\). The algorithm is as follows. Subdivide the manifold \(B\) which is the boundary of \(P\) \[ B=\sum^{m}_{q=1}S_ q=\sum^{m}_{q=1}\langle X^ q_ 1,\ldots,X^ q_ n\rangle. \] Define a matrix \(R(S_ q,F)=(r_{ij})\) where \(r_{ij}=1\) if \(\text{sgn}(f_ i(X^ q_ i))=1\), \(r_{ij}=0\) if \(\text{sgn}(f_ i(X^ q_ i))=- 1\) where we define \(\text{sgn}(0)=1\). Determine \(\text{Par}(R(S_ q,F))\), the parity of \(R(S_ q,F)\). Calculate the sum of \(\text{Par}(R(S_ q,F))\) where \(q\) runs from 1 to \(m\). We can show that \(\deg (F,P,0)=\sum^{m}_{q=1}\text{Par}(R(S_ q,F))\) if the subdivision is small enough. Obviously, the algorithm is combinatorial. In the last section a generalized method of bisection for the solution of nonlinear equations is described, which is based on the algorithm above and the additive principle of mapping degree. The numerical test reports that the algorithm for the calculation of the mapping degree and the generalized method of bisection are efficient for lower-dimensional cases and are not satisfactory for higher-dimensional cases.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3345693)