Methodologies and techniques have been proposed recently for networked system analysis ([3], [4]). A Networked system is a multi-components and multi-functions one; its mission is to supply each user node from at least a source node, through at least a path made up of available edges [2]. In its nominal operating mode, it supplies the whole demand; different degraded operating modes are defined referring to the amount of supplied demand when one or more edges are unavailable. We propose a general approach to assess the Local and Global performances of the system, referring to each user node and to the whole network respectively. We will show that the approaches proposed in literature are consistent with the “enumeration of the system state” and provide a model for networks made up of weighted, unfaultable user nodes and directed, binary edges. The system’s elements (edges) are ranked basing on a metric which is referred to the whole network, by means of an additive importance measures: the Differential Importance Measure (DIM). In order to define a risk-informed measure referred to the Global performance of the system, analytical evidences are provided and an applicative case is developed for a simple network. Keywords: network, importance measure, global performance
Risk significance importance measures for a networked system
LA ROVERE, STEFANO;VESTRUCCI, PAOLO;SPERANDII, MARIA
2008
Abstract
Methodologies and techniques have been proposed recently for networked system analysis ([3], [4]). A Networked system is a multi-components and multi-functions one; its mission is to supply each user node from at least a source node, through at least a path made up of available edges [2]. In its nominal operating mode, it supplies the whole demand; different degraded operating modes are defined referring to the amount of supplied demand when one or more edges are unavailable. We propose a general approach to assess the Local and Global performances of the system, referring to each user node and to the whole network respectively. We will show that the approaches proposed in literature are consistent with the “enumeration of the system state” and provide a model for networks made up of weighted, unfaultable user nodes and directed, binary edges. The system’s elements (edges) are ranked basing on a metric which is referred to the whole network, by means of an additive importance measures: the Differential Importance Measure (DIM). In order to define a risk-informed measure referred to the Global performance of the system, analytical evidences are provided and an applicative case is developed for a simple network. Keywords: network, importance measure, global performanceI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
