A new method for analysis of Multi-State systems based on Multi-valued decision diagram under epistemic uncertainty. (January 2023)

• Record Type: Journal Article
• Authors: Zaitseva, Elena; Levashenko, Vitaly; Rabcan, Jan
• Is Part Of: Reliability engineering & system safety. Volume 229(2023)
• Publication Date: 2023-01
• Subjects: Multi-State System (MSS) -- Fuzzy Decision Tree -- Multi-Valued Decision Diagram -- Epistemic uncertainty -- Availability -- Importance measures; Reliability (Engineering) -- Periodicals; System safety -- Periodicals; Industrial safety -- Periodicals; Fiabilité -- Périodiques; Sécurité des systèmes -- Périodiques; Sécurité du travail -- Périodiques; 620.00452
• DOI: 10.1016/j.ress.2022.108868
• Languages: English
• ISSNs: 0951-8320

Abstract

Highlights: A new method for construction of a mathematical model for reliability analysis. Multi-Valued Decision Diagram construction for Multi-State System. Reliability analysis based on uncertain and incompletely specified data. The system importance analysis based on the proposed method. Abstract: The development of the system's model is an important step in reliability analysis. The system's model in reliability analysis similar to other knowledge areas is used as an explanatory and research tool. In reliability analysis, this model is mathematical and allows us to determine quantitative characteristics to evaluate the behavior of the original system. Typically such a mathematical model approximates the behavior of the initial system and has some uncertainty. However, this uncertainty can increase significantly if the initial data for building the model is uncertain and incompletely specified. The uncertainty of the mathematical model also causes incorrect estimates of the system behavior and its reliability. Therefore, it is important to develop methods that allow us to take into account the uncertain nature of the initial data, first of all, take into account epistemic uncertainty in initial data. A new method for the development of a mathematical model of a Multi-State System (MSS) in the form of a Multi-Valued Decision Diagram based on incompletely specified and uncertain data is proposed. In other words, the proposed method takes into account the epistemic uncertainty of the initial data. The specific of this method is the use of Data Mining based classification procedures, in particular, the Fuzzy Decision Trees.