This paper develops a model for predicting the failure time of a wide class of weighted k- out-of-n reliability systems. To this aim, we adopt a rational expectation-type approach by arti cially creating an information set based on the observation of a collection of systems of the same class -- the catalog. Speci cally, we state the connection between a synthetic statistical measure of the survived components' weights and the failure time of the systems. In detail, we follow the evolution of the systems in the catalog from the starting point to their failure { obtained after the failure of some of their components. Then, we store the couples given by the measure of the survived components and the failure time. Finally, we employ such couples for having a prediction of the failure times of a set of new systems { the in-vivo systems { conditioned on the speci c values of the considered statistical measure. We test di erent statistical measures for predicting the failure time of the in-vivo systems. As a result, we give insights on the statistical measure which is more effective in contributing to providing a reliable estimation of the systems' failure time. A discussion on the initial distribution of the weights is also carried out.
Rational expectations as a tool for predicting failure of weighted k-out-of-n reliability systems / Andersen, J. -V.; Cerqueti, R.; Riccioni, J.. - In: ANNALS OF OPERATIONS RESEARCH. - ISSN 1572-9338. - (2023).
Rational expectations as a tool for predicting failure of weighted k-out-of-n reliability systems
R. Cerqueti;J. Riccioni
2023
Abstract
This paper develops a model for predicting the failure time of a wide class of weighted k- out-of-n reliability systems. To this aim, we adopt a rational expectation-type approach by arti cially creating an information set based on the observation of a collection of systems of the same class -- the catalog. Speci cally, we state the connection between a synthetic statistical measure of the survived components' weights and the failure time of the systems. In detail, we follow the evolution of the systems in the catalog from the starting point to their failure { obtained after the failure of some of their components. Then, we store the couples given by the measure of the survived components and the failure time. Finally, we employ such couples for having a prediction of the failure times of a set of new systems { the in-vivo systems { conditioned on the speci c values of the considered statistical measure. We test di erent statistical measures for predicting the failure time of the in-vivo systems. As a result, we give insights on the statistical measure which is more effective in contributing to providing a reliable estimation of the systems' failure time. A discussion on the initial distribution of the weights is also carried out.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.