Representing results of V&V-activities in fuzzy decision analysis value trees
2005 (English)In: European Simulation Interoperability Workshop 2005, 2005, 74-87 p.Conference paper (Refereed)
The process of accrediting a model can roughly be divided into first gathering and evaluating results of conducted V&V activities, and then aggregating these results into a total value: The accreditation decision. Especially the second part is not trivial, because results of V&V activities are quite manifold and at first sight not comparable with each other. Consider for example the evaluation of subject matter expert statements in contrast to real numbers as an outcome of statistical tests. Classical decision analysis, especially value tree analysis, has been implemented to support the accreditation decision by using a structured, scientifically justified approach. However, traditional value tree analysis is not flawless and exhibits deficiencies, as there are for example: The impossibility to distinguish between compensatory resp. non compensatory attribute values, the impossibility to not only quantify the knowledge about a model, but also the ignorance ("How much is known to be not known about the model?"), and the insufficient flexibility of using real or natural numbers to represent subject matter expert statements. In a previous paper (05S-SIW-042) the usage of Fuzzy Multi Attribute Decision Making, a branch of Decision Analysis incorporating Fuzzy Set Theory, was proposed in order to overcome the above mentioned deficiencies. Also a list of tasks was depicted, what needs to be done to enrich traditional value tree analysis by fuzzy set theoretic concepts. The very first task is the design of the attribute values, i. e. how leaves of a fuzzy value tree allegorizing results of conducted V&V activities look like, and this is the substance of this paper: The design of attribute values in a fuzzy value tree analysis in order to support the accreditation decision. The focal point is the representation of subject matter expert statements not by real or natural numbers as in traditional value tree analysis, but by fuzzy sets. After a short recapitulation of fuzzy set theory basics, the different designs to model value functions with the help of fuzzy sets are depicted. Advantages and disadvantages of each design are enumerated, and finally one specific concept, regarding the particularities of modeling the results of V&V activities, and the deficiencies of traditional value tree analysis mentioned above, is put forward.
Place, publisher, year, edition, pages
2005. 74-87 p.
Accreditation, Decision theory
IdentifiersURN: urn:nbn:se:kth:diva-148509ScopusID: 2-s2.0-84867561089ISBN: 978-162276141-8OAI: oai:DiVA.org:kth-148509DiVA: diva2:741081
European Simulation Interoperability Workshop 2005; Toulouse, France, 27-29 June, 2005
QC 201408272014-08-272014-08-082014-08-27Bibliographically approved