Autoren: Slim, Nadeau
Kategorie: Technikfolgen und -abschätzung
In recent years, considerable research efforts in safety management were directed at proposing innovative methodological frameworks to address the complexity of modern sociotechnical systems. The significance of results in such endeavors, whether quantitative or qualitative, relies largely on the quality of input data and the validity of the implemented methods to model such systems. To provide more objective and valid results, new protocols and tools for data processing are needed as well. An interesting data-mining tool for computing with incomplete and uncertain information is Rough Set Theory (RST). In this study, we propose the application of RST to generate comprehensible IF-THEN rule bases from data tables for classifying outcomes within the framework of the Functional Resonance Analysis Method (FRAM). For this purpose, a list of Common Performance Conditions (CPC) was used as performance indicators to construct rough decision systems. The generated rule base was applied in combination with our fuzzy FRAM prototyping model in a realistic case study simulating the working environment of aircraft deicing operations. The obtained results showed high accuracy and the generated rule base for each function was reduced in size significantly, which allowed for a more efficient simulation. The demonstration here presents an ideal scenario with ideal data sets, which resulted in maximum accuracy. The model however still requires further optimization and validation using expert’s input data in future applications. Such an approach can allow for more intersubjective and comprehensible results, which would result in a better decision-making process and consequently a more resilient systemic performance.
Zeitfenster der Veranstaltung (1)