Local search with groups of step sizes.

No Thumbnail Available
Date
2021
Authors
Costa, Rodolfo Ayala Lopes
Freitas, Alan Robert Resende de
Silva, Rodrigo César Pedrosa
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Local search methods for continuous optimization problems tend to be sensitive to the choice of step sizes in their search directions. This paper presents the Local Search with Groups of Step Sizes (LSGSS) method, a derivative-free method that reactively updates groups of promising step sizes for each problem coordinate. The experiments demonstrate LSGSS could find the best solutions for each large-scale benchmark problem when compared to classical methods.
Description
Keywords
Continuous optimization, Derivative-free local search
Citation
COSTA, R. A. L.; FREITAS, A. R. R. de; SILVA, R. C. P. Local search with groups of step sizes. Operations Research Letters, v. 49, p. 385-392, 2021. Disponível em: <https://www.sciencedirect.com/science/article/abs/pii/S016763772100050X>. Acesso em: 25 ago. 2021.