Polynomial acceleration of Monte-Carlo global search

Document Type

Conference Proceeding

Publication Date

1-1-1999

Abstract

In this paper we describe a class of algorithms for approximating the global minimum of a function defined on a subset of d-dimensional Euclidean space. The algorithms are based on adaptively composing a number of simple Monte Carlo searches, and use a memory of a fixed finite number of observations. Within the class of algorithms it is possible to obtain arbitrary polynomial speedup in the asymptotic convergence rate compared with simple Monte Carlo.

Identifier

0033339658 (Scopus)

Publication Title

Winter Simulation Conference Proceedings

External Full Text Location

https://doi.org/10.1145/324138.324454

ISSN

02750708

First Page

673

Last Page

677

Volume

1

This document is currently not available here.

Share

COinS