Semiparametric statistical inference in global random search
AA Zhigljavsky - Acta Applicandae Mathematica, 1993 - Springer
Acta Applicandae Mathematica, 1993•Springer
The multivariate multiextremal optimization problem is considered. Various statistical
procedures based on the use of the asymptotic theory of extreme order statistics are
thoroughly described. These procedures are used to infer about the maximal value of a
function by its values at random points. A class of global random search methods underlying
the procedures is considered. These methods generalize the well-known branch and bound
methods. The article is mainly of a survey nature. It also contains new results.
procedures based on the use of the asymptotic theory of extreme order statistics are
thoroughly described. These procedures are used to infer about the maximal value of a
function by its values at random points. A class of global random search methods underlying
the procedures is considered. These methods generalize the well-known branch and bound
methods. The article is mainly of a survey nature. It also contains new results.
Abstract
The multivariate multiextremal optimization problem is considered. Various statistical procedures based on the use of the asymptotic theory of extreme order statistics are thoroughly described. These procedures are used to infer about the maximal value of a function by its values at random points. A class of global random search methods underlying the procedures is considered. These methods generalize the well-known branch and bound methods. The article is mainly of a survey nature. It also contains new results.
Springer