Prof. Yaroslav D. Sergeyev and PhD Dmitry E. Kvasov, academics from the Software Department of Lobachevsky University (UNN) won the 2014 Journal of Global Optimization Best Paper Award for the paper entitled “Globally-biased DISIMPL algorithm for expensive global optimization” (July 2014, Volume 59, Pages 545-567), co-authored with Remigijus Paulavicius and Julius Zilinskas, researchers from the Institute of Mathematics and Informatics of Vilnius University.

Springer’s Journal of Global Optimization is an international journal on theoretical, computational and applied aspects of global optimization, numerical experiments and applications in engineering, management and the sciences (a journal impact factor of 1.355). Its Editorial Board includes prominent scientists from the most famous universities in the world.

Abstract

Direct-type global optimization algorithms often spend an excessive number of function evaluations on problems with many local optima exploring suboptimal local minima, thereby delaying discovery of the global minimum. In this paper, a globally-biased simplicial partition Disimpl algorithm for global optimization of expensive Lipschitz continuous functions with an unknown Lipschitz constant is proposed. A scheme for an adaptive balancing of local and global information during the search is introduced, implemented, experimentally investigated, and compared with the well-known Direct and Directl methods. Extensive numerical experiments executed on 800 multidimensional multiextremal test functions show a promising performance of the new acceleration technique with respect to competitors.

Read the paper online