×

zbMATH — the first resource for mathematics

On the identification of local minimizers in inertia-controlling methods for quadratic programming. (English) Zbl 0737.65047
For the general quadratic nonconvex programming problem: minimize \(\varphi(x)=c^ Tx+{1\over 2} x^ T Hx\) subject to \(Ax\geq\beta\), where the Hessian matrix is symmetric and \(A\) is an \(m\times n\) matrix, the verification of optimality is discussed within the context of an inertia- controlling method. A computational method is derived that will attempt to determine if a dead point is a local minimizer.

MSC:
65K05 Numerical mathematical programming methods
90C20 Quadratic programming
PDF BibTeX XML Cite
Full Text: DOI