Volume 15, No 3, 2008, P. 43-57

UDC 519.87
T. V. Levanova, A. S. Fedorenko
Variable neighborhood search for two-stage facility location problem

Abstract:
In this paper, a heuristic algorithm of variable neighborhood search (VNS) is developed for solving two-stage facility location problem. In the VNS framework, a suitable set of neighborhoods and the order of their exploration are proposed. Besides the well-known neighborhoods, a new type of neighborhood is used. Computational experiments on benchmark instances show that in most testing instances the method gives better solutions than the simulated annealing algorithm. 
Tabl. 5, illustr. 3, bibl. 19.

Keywords: discrete optimization, two-stage facility location problem, variable neighborhood search.

Levanova Tatyana Valentinovna ¹
Fedorenko Anatoliy Sergeevich ¹
1. Omsk Branch of Sobolev Institute of Mathematics, SB RAS,
13, Pevtsov str., 644099, Omsk, Russia
e-mail: tlevanova@ofim.oscsbras.ru, fas.omsk@mail.ru