Univariate box-constrained Lipschitz global optimization problems are considered in this contribution. Geometric and information statistical approaches are presented. The novel powerful local tuning and local improvement techniques are described in the contribution as well as the traditional ways to estimate the Lipschitz constant. The advantages of the presented local tuning and local improvement techniques are demonstrated using the operational characteristics approach for comparing deterministic global optimization algorithms on the class of 100 widely used test functions.
Acceleration techniques in the univariate Lipschitz global optimization
Yaroslav D. Sergeyev, Dmitri E. Kvasov, Marat S. Mukhametzhanov, Angela De Franco; Acceleration techniques in the univariate Lipschitz global optimization. AIP Conf. Proc. 20 October 2016; 1776 (1): 090051. https://doi.org/10.1063/1.4965415
Download citation file: