Publication:
Generalized reduced gradient algorithm for a local optimum of a cost function built through simulation
Generalized reduced gradient algorithm for a local optimum of a cost function built through simulation
dc.contributor.advisor | Bartolomei-Suárez, Sonia M. | |
dc.contributor.author | Seijo-Vidal, Roberto L. | |
dc.contributor.college | College of Engineering | en_US |
dc.contributor.committee | Irizarry, María | |
dc.contributor.committee | Pagán Parés, Omell | |
dc.contributor.department | Department of Industrial Engineering | en_US |
dc.contributor.representative | Toledo, Freya | |
dc.date.accessioned | 2018-11-28T17:29:11Z | |
dc.date.available | 2018-11-28T17:29:11Z | |
dc.date.issued | 2004 | |
dc.description.abstract | This project is the continuation of a previous simulation study based on a trial and error approach that pretended to find a better system. This new phase pursued a scientific approach for the simulation study in order to identify the best alternative: sensitivity analysis, design of experiments, regression analysis for metamodeling purposes, and optimization. Typical simulation optimization methods were not of practical value for this application. An optimization tool based on mathematical programming was developed using Microsoft’s Excel Solver. The tool was validated in terms of the metamodels accuracy and the capacity to find a local optimum within the search region. It was concluded that additional experimental designs were needed in order to find the global optimum. Nevertheless, the tool was valid for the practical application of this project. Finally, it was also concluded that the scientific approach rendered better results than the trial and error approach. | en_US |
dc.description.abstract | Este proyecto es la continuación de un estudio de simulación que pretendía resolver un problema de aplicación práctica siguiendo una metodología de prueba y error. Esta nueva fase perseguía encontrar la mejor alternativa siguiendo una metodología científica: análisis de sensitividad, diseño de experimentos, metamodelos y optimización. Se determinó que los métodos de optimización comúnmente utilizados en simulación no eran prácticos para este problema. Se desarrolló una herramienta de optimización basada en programación matemática con el uso de “Microsoft Excel Solver”. La herramienta fue validada en términos de la precisión de los metamodelos y la capacidad para encontrar el óptimo dentro de la región de búsqueda. Se concluyó que para determinar el óptimo verdadero, era necesario realizar experimentos adicionales. Sin embargo, se demostró que la herramienta era válida para el problema que se pretendía resolver. Por último, se demostró que el método científico rindió mejores beneficios que el método de prueba y error. | en_US |
dc.description.graduationYear | 2004 | en_US |
dc.description.sponsorship | Industrial Engineering Department | en_US |
dc.identifier.uri | https://hdl.handle.net/20.500.11801/1544 | |
dc.language.iso | English | en_US |
dc.rights.holder | (c)2004 Roberto L. Seijo Vidal | en_US |
dc.rights.license | All rights reserved | en_US |
dc.subject | Simulation study | en |
dc.subject | Optimization tool | en |
dc.subject | Local optimum | en |
dc.title | Generalized reduced gradient algorithm for a local optimum of a cost function built through simulation | en_US |
dc.type | Thesis | en_US |
dspace.entity.type | Publication | |
thesis.degree.discipline | Industrial Engineering | en_US |
thesis.degree.level | M.S. | en_US |
Files
Original bundle
1 - 1 of 1
- Name:
- ININ_SeijoVidalRL_2004.pdf
- Size:
- 3.01 MB
- Format:
- Adobe Portable Document Format
- Description: