My SciELO
Services on Demand
Article
Indicators
- Cited by SciELO
Related links
- Similars in SciELO
Share
Ingeniería Mecánica
On-line version ISSN 1815-5944
Abstract
ANTONIO-MAGAIA, Nabote; ALEMAN-ROMERO, Isidro Luis; ARZOLA-RUIZ, José and MARTINEZ-VALDES, Osmel. Posynomial fitting combiding Particle Swarm and Nonlinear Programming Methods. Ingeniería Mecánica [online]. 2017, vol.20, n.2, pp. 83-90. ISSN 1815-5944.
The performance of deterministic methods of nonlinear programming is highly sensitive to the selection of the initial approximation, particularly when it comes to obtaining an overall optimum, as required in the posyinomial fitting. The present work aimed to generate the initial approximation through of particle swarm metaheuristic, to improve the performance of non-linear programming methods in posyinomial fitting. From the non-parametric tests performed for the comparison of the results obtained in the considered problems, it was observed that the differences between the results of the posyinomial fitting, with and without the application of the improvement proposal, are significant. In addition, it was found that the improvement methodology developed allows to obtain a good performance of the non-linear programming methods in the posynomial fitting, which results in the obtaining of quality posynomial models.
Keywords : posynomial fitting; nonlinear programming; particle swam optimization; geometric programming.