An artificial fish swarm algorithm based hyperbolic augmented Lagrangian method

• Author(s): Costa, M. Fernanda P. ; Rocha, Ana Maria A. C. ; Fernandes, Edite Manuela da G. P.
• Date: 2013
• Persistent ID: http://hdl.handle.net/1822/26658
• Origin: RepositóriUM - Universidade do Minho
• Subject(s): Augmented Lagrangian; Hyperbolic penalty; Artificial fish swarm; Stochastic convergence
• Document Type: Preprint
• Language: English

Description

Documento submetido para revisão pelo pares a publicar em Journal of computational and applied mathematics. ISSN 0377-0427. Versão "In Press, Corrected Proof" disponível em http://www.sciencedirect.com/science/article/pii/S0377042713004226 This paper aims to present a hyperbolic augmented Lagrangian (HAL) framework with guaranteed convergence to an ϵ-global minimizer of a constrained nonlinear optimization problem. The bound constrained subproblems that emerge at each iteration k of the framework are solved by an improved artificial fish swarm algorithm. Convergence to an ϵk-global minimizer of the HAL function is guaranteed with probability one, where ϵk→ϵ as k→∞. Preliminary numerical experiments show that the proposed paradigm compares favorably with other penalty-type methods.