Javascript must be enabled for the correct page display

Estimating Values of Complex Dependencies in COVAMOF

Businge, J. (2006) Estimating Values of Complex Dependencies in COVAMOF. Master's Thesis / Essay, Computing Science.

[img]
Preview
Text
Infor_Ma_2006_JBusinge.CV.pdf - Published Version

Download (2MB) | Preview

Abstract

Complex dependencies in COVAMOF (Configuration of Industrial Product Families Variability Modeling Framework) are dependencies that are affected by a large number of variation points. In most cases, they may not be stated formally, but have some informal character, e.g. "the combination of parameter settings will have a negative effect on the performance of the overall software system". In this thesis, we use mathematical algorithms of interpolation and extrapolation to try to formalize the determination of complex dependency values. We specifically look at complex dependency values that are affected by the variations in two and three value bindings. For dependency values that are affected by variations of two value bindings, we compare estimates obtained by interpolation/extrapolation algorithms like 3-term Level plane, 4-term Bilinear, 8-term Biquadratic, 9-term Biquadratic, and 16-term Bicubic equations. For dependency values that are affected by three value bindings, we compare estimates obtained by techniques like 7-term Trilinear, 8-term Trilinear, 27-term Triquadratic, and 64-term Tricubic. Our findings from the data set used show us that there is no dominating algorithm in producing better estimates. Lower order algorithms like 3-term Level plane and 4-term Bilinear for 2 dimensions and 7-term Trilinear and 8-term Trilinear seems a well thought alternative since they use the fewest data points and have the smallest execution time in their computations.

Item Type: Thesis (Master's Thesis / Essay)
Degree programme: Computing Science
Thesis type: Master's Thesis / Essay
Language: English
Date Deposited: 15 Feb 2018 07:30
Last Modified: 15 Feb 2018 07:30
URI: http://fse.studenttheses.ub.rug.nl/id/eprint/8953

Actions (login required)

View Item View Item