R&D project performance evaluation with multiple and interdependent criteria

Tohumcu, Zeynep
In this study, an Analytic Network Process (ANP) and Data Envelopment Analysis (DEA) based approach was developed in order to measure the performance of customer-based Research and Development projects being executed in TÜB TAKSAGE, Defense Research and Development Institute, under the Scientific and Technological Research Council of Turkey. In order to evaluate project performance, many criteria, containing various subcriteria were determined. In order to handle the interdependencies among the criteria and the sub-criteria, ANP was used. The ANP model generated in this study is a hybrid model consisting of both a hierarchy and a network. The pairwise comparison matrices that were built up for defining the importance and influences of the criteria/sub-criteria in the ANP model were formed as interval judgments from a group decision making process, based on data obtained from a questionnaire conducted among the experts in the Institute. From the interval pairwise comparison matrices, weight intervals for the sub-criteria were determined and these bounds were used as assurance region constraints in a super-efficiency DEA model, through which the project ranking was obtained. Taking into consideration that there may occur some missing values in some projects for some of the sub-criteria, the superefficiency DEA model was extended to handle missing data. The model was applied to a real case study on performance evaluation of the ongoing customer-based projects in the Institute. For comparison purposes, the case study was also solved by two other approaches.
Citation Formats
Z. Tohumcu, “R&D project performance evaluation with multiple and interdependent criteria,” M.S. - Master of Science, Middle East Technical University, 2007.