Risk assessment and management of civil infrastructure networks: a systems approach - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 437

￿ Intra-cluster relationships: These are relationships among actual network

nodes within a specifi c fi ctitious node, which are represented by the
sub-graph G [ V i ( lv ) , E ( V i ( lv ) )] containing only the elements from the actual
network that lie within V i ( lv ) .

￿ Inter-level relationships: These are relationships between fi ctitious nodes

at adjacent levels of the hierarchy and denote how fi ctitious nodes
(clusters) are embedded into others.

Each level in the hierarchy provides a different degree of detail. Therefore,
a problem can be signifi cantly simplifi ed and emergent properties can
be detected (Gómez et al. , 2011a); in fact, the system should not be taken
to a level of detail beyond which the relevance of information is more
valuable than its specifi city. The main advantages of using fi ctitious net-
works are:

￿ Enhanced insight into emergent properties of the network at different
levels of resolution, which is important for risk analysis and decision
making (e.g., detecting critical elements).
￿ The computational burden of network analyses can be substantially
reduced by balancing the specifi city and relevance of information.

17.4 Risk assessment of infrastructure networks
Risk is related to the effect of uncertainty on the performance of a system
of interest. Moreover, hazards to the system may lead to vulnerability when
compromising critical elements. Risk analysis is essential for decision
making: rational and structured analyses of available information provide
key evidence for effective decisions. Furthermore, this type of analyses is
closely tied to the socioeconomic background of involved individuals and
organizations.
Risk assessment requires the identifi cation and valuation of future sce-
narios; then, all results are estimates of probabilities of occurrence of events
and their consequences. Managing uncertainty is vital in this process
(Knight, 2009), which includes lack of knowledge about the problem and
randomness. Even if all scenarios are known, the combinatorial problem
would be excessively large and diffi cult to evaluate. Finally, risk assessment
is essential for risk management tasks, such as risk mitigation/reduction. The
following sections discuss the use of the proposed systems approach and
hierarchical model for the following problems:

￿ reliability assessment;
￿ vulnerability analysis;
￿ damage propagation analysis.