Separable cubic modeling and a trust-region strategy for unconstrained minimization with impact in global optimization

Autor(es)

Martinez, J. M. ^[1] ; Raydan, M. ^[2]

Afiliação do(s) autor(es)

^[1] Univ Estadual Campinas, IMECC UNICAMP, Dept Appl Math, BR-13083859 Campinas, SP - Brazil

^[2] Univ Simon Bolivar, Dept Comp Cient & Estadist, Caracas 1080 - Venezuela

• Tipo de documento: Artigo Científico
• Fonte: Journal of Global Optimization; v. 63, n. 2, p. 319-342, OCT 2015.
• Citações Web of Science: 3

Resumo

A separable cubic model, for smooth unconstrained minimization, is proposed and evaluated. The cubic model uses some novel secant-type choices for the parameters in the cubic terms. A suitable hard-case-free trust-region strategy that takes advantage of the separable cubic modeling is also presented. For the convergence analysis of our specialized trust region strategy we present as a general framework a model -order trust region algorithm with variable metric and we prove its convergence to -stationary points. Some preliminary numerical examples are also presented to illustrate the tendency of the specialized trust region algorithm, when combined with our cubic modeling, to escape from local minimizers. (AU)

• Processo FAPESP: 13/07375-0 - CeMEAI - Centro de Ciências Matemáticas Aplicadas à Indústria
• Beneficiário: Francisco Louzada Neto
• Modalidade de apoio: Auxílio à Pesquisa - Centros de Pesquisa, Inovação e Difusão - CEPIDs
• Processo FAPESP: 13/05475-7 - Métodos computacionais de otimização
• Beneficiário: Sandra Augusta Santos
• Modalidade de apoio: Auxílio à Pesquisa - Temático