TY - GEN
T1 - Estimating probability of failure of a complex system based on inexact information about subsystems and components, with potential applications to aircraft maintenance
AU - Kreinovich, Vladik
AU - Jacob, Christelle
AU - Dubois, Didier
AU - Cardoso, Janette
AU - Ceberio, Martine
AU - Batyrshin, Ildar
PY - 2011
Y1 - 2011
N2 - In many real-life applications (e.g., in aircraft maintenance), we need to estimate the probability of failure of a complex system (such as an aircraft as a whole or one of its subsystems). Complex systems are usually built with redundancy allowing them to withstand the failure of a small number of components. In this paper, we assume that we know the structure of the system, and, as a result, for each possible set of failed components, we can tell whether this set will lead to a system failure. For each component A, we know the probability P(A) of its failure with some uncertainty: e.g., we know the lower and upper bounds and for this probability. Usually, it is assumed that failures of different components are independent events. Our objective is to use all this information to estimate the probability of failure of the entire the complex system. In this paper, we describe a new efficient method for such estimation based on Cauchy deviates.
AB - In many real-life applications (e.g., in aircraft maintenance), we need to estimate the probability of failure of a complex system (such as an aircraft as a whole or one of its subsystems). Complex systems are usually built with redundancy allowing them to withstand the failure of a small number of components. In this paper, we assume that we know the structure of the system, and, as a result, for each possible set of failed components, we can tell whether this set will lead to a system failure. For each component A, we know the probability P(A) of its failure with some uncertainty: e.g., we know the lower and upper bounds and for this probability. Usually, it is assumed that failures of different components are independent events. Our objective is to use all this information to estimate the probability of failure of the entire the complex system. In this paper, we describe a new efficient method for such estimation based on Cauchy deviates.
KW - complex system
KW - interval uncertainty
KW - probability of failure
UR - http://www.scopus.com/inward/record.url?scp=82555177332&partnerID=8YFLogxK
U2 - 10.1007/978-3-642-25330-0_7
DO - 10.1007/978-3-642-25330-0_7
M3 - Contribución a la conferencia
SN - 9783642253294
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 70
EP - 81
BT - Advances in Soft Computing - 10th Mexican International Conference on Artificial Intelligence, MICAI 2011, Proceedings
PB - Springer Verlag
T2 - 10th Mexican International Conference on Artificial Intelligence, MICAI 2011
Y2 - 26 November 2011 through 4 December 2011
ER -