Global optimization using q-gradients

•Two global optimization methods based on the q-gradient are developed: q-G and q-CG.•q-G and q-CG are q-analogs of the steepest descent and conjugate gradient methods.•Convergence of q-G and q-CG with Gaussian perturbations is proved.•The methods are compared with their classical versions and with...

Full description

Saved in:
Bibliographic Details
Published in:European journal of operational research 2016-06, Vol.251 (3), p.727-738
Main Authors: Gouvêa, Érica J.C., Regis, Rommel G., Soterroni, Aline C., Scarabello, Marluce C., Ramos, Fernando M.
Format: Article
Language:English
Subjects:
Convergence
Derivatives
Finite difference method
Gaussian
Global optimization
Heuristic
Mathematical analysis
Mathematical problems
Metaheuristics
Methods
Optimization
Optimization techniques
Probability
q-calculus
q-gradient vector
Searching
Studies
Vectors (mathematics)
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Two global optimization methods based on the q-gradient are developed: q-G and q-CG.•q-G and q-CG are q-analogs of the steepest descent and conjugate gradient methods.•Convergence of q-G and q-CG with Gaussian perturbations is proved.•The methods are compared with their classical versions and with other alternatives.•q-G and q-CG are very competitive with the alternative methods on multimodal problems. The q-gradient vector is a generalization of the gradient vector based on the q-derivative. We present two global optimization methods that do not require ordinary derivatives: a q-analog of the Steepest Descent method called the q-G method and a q-analog of the Conjugate Gradient method called the q-CG method. Both q-G and q-CG are reduced to their classical versions when q equals 1. These methods are implemented in such a way that the search process gradually shifts from global in the beginning to almost local search in the end. Moreover, Gaussian perturbations are used in some iterations to guarantee the convergence of the methods to the global minimum in a probabilistic sense. We compare q-G and q-CG with their classical versions and with other methods, including CMA-ES, a variant of Controlled Random Search, and an interior point method that uses finite-difference derivatives, on 27 well-known test problems. In general, the q-G and q-CG methods are very promising and competitive, especially when applied to multimodal problems.
ISSN:0377-2217
1872-6860
DOI:10.1016/j.ejor.2016.01.001