Global minimization of a quadratic functional: neural network approach

12/24/2004

∙

by   L. B. Litinskii, et al.

∙

0

∙

share

The problem of finding out the global minimum of a multiextremal functional is discussed. One frequently faces with such a functional in various applications. We propose a procedure, which depends on the dimensionality of the problem polynomially. In our approach we use the eigenvalues and eigenvectors of the connection matrix.