Multistart Hooke and Jeeves filter method for mixed variable optimization

• Autor(es): Fernandes, Florbela P. ; Costa, M. Fernanda P. ; Fernandes, Edite Manuela da G. P. ; Rocha, Ana Maria A. C.
• Data: 2013
• Identificador Persistente: http://hdl.handle.net/1822/26678
• Origem: RepositóriUM - Universidade do Minho
• Assunto(s): Global optimization; Multistart; Hooke and Jeeves; Filter method
• Tipo de Documento: Documento de conferência
• Idioma: Inglês

Descrição

AIP Conference Proceedings, vol. 1558 In this study, we propose an extended version of the Hooke and Jeeves algorithm that uses a simple heuristic to handle integer and/or binary variables and a filter set methodology to handle constraints. This proposal is integrated into a multistart method as a local solver and it is repeatedly called in order to compute different optimal solutions. Then, the best of all stored optimal solutions is selected as the global optimum. The performance of the new method is tested on benchmark problems. Its effectiveness is emphasized by a comparison with other well-known stochastic solvers.