scholarly journals Sensitivity Analysis for Two-Level Value Functions with Applications to Bilevel Programming

2012 ◽  
Vol 22 (4) ◽  
pp. 1309-1343 ◽  
Author(s):  
S. Dempe ◽  
B. S. Mordukhovich ◽  
A. B. Zemkoho
2017 ◽  
Vol 117 (7) ◽  
pp. 1362-1375 ◽  
Author(s):  
Carolina Lino Martins ◽  
Hipólito Marcelo Losada López ◽  
Adiel Teixeira de Almeida ◽  
Jonatas Araújo Almeida ◽  
Mirian Batista de Oliveira Bortoluzzi

Purpose The purpose of this paper is to analyze the impacts of Portfolio size effect due to scaling issues in the outcome obtained in a project portfolio selection for an electricity company in Brazil, focusing on improving business strategic performance. Design/methodology/approach The study uses a web-based decision support system (DSS), in which scaling issues are considered, incorporating results of previous work. The study evaluates 32 projects from the electricity company and compared the possible results when considering different scales. Additionally, a sensitivity analysis was conducted to analyze the robustness of the case, using the web-based DSS. Findings The results for an interval scale context showed a portfolio with 21 projects, contrasting with the correct solution of a portfolio containing 23 projects. The latter is related to a ratio scale context, with the proper transformation of weights, which was found to be robust with a sensitivity analysis using Monte Carlo simulation. This demonstrates that only appropriate models for selecting projects can improve the contribution to the company’s permanent strategies of increasing productivity, considering its constraints to achieve optimal results. Originality/value Additive value functions approach imposes certain requirements on the measurement scales used for the items in a portfolio that should not be ignored, once they have significant impact on the general portfolio results, which are directly related to the business strategic performance and the facilities of doing that with a web-based DSS.


Author(s):  
Stephane Ross ◽  
Masoumeh Izadi ◽  
Mark Mercer ◽  
David Buckeridge

Sign in / Sign up

Export Citation Format

Share Document