Content uploaded by Silvia De Angeli
Author content
All content in this area was uploaded by Silvia De Angeli on Aug 19, 2022
Content may be subject to copyright.
2nd Scientific International Conference on CBRNe SICC Series –10 –12 December 2020
Pasino A. (1), Battista U. (2), De Angeli S. (3), Ottonello D. (2), Clematis A. (1)
(1) CNR  IMATI, Via De Marini 6, 16149 Genoa, Italy; (2) Stam S.r.l., Via Pareto 8 AR, 16129 Genoa, Italy; (3) Department of
Civil, Chemical and Environmental Engineering, University of Genoa, 16145, Genoa, Italy.
Alessandro Pasino (Poster Presenter) –CNRIMATI
Via De Marini 6  16149 Genoa, Italy
3349016802 –alessandro.pasino@ge.imati.cnr.it
A review of single and multihazard risk assessment approaches for critical
infrastructures protection
Data Collection MultiHazard Interactions
Methodologies
MATHEMATICAL AND STATISTICAL METHODS MACHINE LEARNING TECHNIQUES GRAPHS AND NETWORKS APPROACH
•
Bayesian Belief Network (BBN)
•
Logarithmic regression
•
Early warning index calculation
•
Artificial Neural Network (ANN)
•
Support Vector Machine (SVM)
•
Boosted Regression Tree (BRT)
•
Generalized Additive Model (GAM)
•
Genetic Algorithm
Rule
Set Production (GARP)
•
Quick Unbiased Efficient
Statistical Tree (QUEST)
•
Game theory technique
•
Classical complex networks methodology
•
Multilevel complex network formulation
Pros &Cons
MATHEMATICAL AND STATISTICAL METHODS MACHINE LEARNING TECHNIQUES GRAPHS AND NETWORKS APPROACH
•
BBNs work well with few data and capture
the
dependencies among variables
•
The logarithmic regression gives good results even
with
poor data accompanied by expert opinions
•
The early warning index highlights well the
correlations
among variables
•
ANNs find amazing results also with poor data and
are
very good for evaluating correlations
•
SVM,BRT and GAM sometimes show good
performance
for the hazards predicted
•
GARP and QUEST require a little quantity of data and
are
very quick methodologies
•
Game theory models are easy to build and can be
used
for more hazards together
•
Classical complex networks require few data
•
Multilevel complex networks perform well both
with
small and with large data sets and are able to deal
with
interdependencies among different hazards
•
The more complex a BBN, the more data are required
•
The logarithmic regression does not capture well
the
interdependencies among variables
•
The early warning index calculation requires lots of data
•
ANNs perform badly with too many variables and
are
able to predict one hazard at a time
•
SVM,BRT and GAM can consider one hazard at a
time
and can show weak results for the hazards predicted
•
GARP and QUEST deal with one hazard at a time
•
Game theory needs expert judgement
•
Classical complex networks work with one hazard at
a
time
•
Multilevel complex networks are used only with
time
series with the same number of elements
CRITICAL INFRASTRCUTURE (CI)
A
system which must be
constantly
monitored
because its destruction
or
interruption
of service brings to
a
weakening
of the efficiency of a city
or
of
an entire country
NATURAL HAZARDS
Dangers
whose origin becomes from
nature.
Examples
of natural hazards are
hurricanes,
floods,
landslides, etc.
MANMADE HAZARDS
Dangers
whose origin is
anthropogenic.
Examples
of manmade hazards are
terroristic
attacks,
crimes, etc.
is subject to
brings to
POOR DATA
Data
must be integrated thanks
to
expert
judgment.
This
subjective knowledge
is
supported
by:
•
Bayes’ theorem
•
Probability bound analysis
RICH DATA
Data
can be objectively
stored
thanks
to triplets:
•
th scenario;
•
probability of 
th scenario;
•
effect of the th scenario.
INDEPENDENT EVENTS
Considering
the events as independent
is
less
complex, but brings to
underestimate
risk
. The method used is the
total
exceedance
probability:
CORRELATED EVENTS
Considering
the events as correlated
is
more
accurate, but the cases to
consider
are
a lot:
•
Unidirectional vs.bidirectional
hazards
•
Triggering vs.increased probability
vs.
catalysis or impedance
Conditional
probability is then used
split into
Acknowledgments: This work has been cofinaced by
Programma Operativo Por FSE Regione Liguria 20142020
Under grant GRISK code RLOF18ASSRIC/70/
Depending on the quantity of data available and on the type of interactions, it is possible choosing
the best methodology to use for the single or multihazard problem faced
References
[
1] Chatterjee, S., & Abkowitz, M. D. (2011). A methodology for modeling regional terrorism risk. Risk Analysis:
An
International
Journal, 31(7), 11331140.
[
2] Liu, Z., Nadim, F., GarciaAristizabal, A., Mignan, A., Fleming, K., & Luna, B. Q. (2015). A threelevel
framework
for
multirisk assessment. Georisk: Assessment and management of risk for engineered systems and geohazards
,
9
(2), 5974.
[
3] Major, J. A. (2002). Advanced techniques for modeling terrorism risk. Journal of Risk Finance, 4(1), 1524.
[
4] Ferson, S., & Ginzburg, L. R. (1996). Different methods are needed to propagate ignorance and variability
.
Reliability
Engineering & System Safety, 54(23), 133144.
[
5] Daskalaki, E., Spiliotis, K., Siettos, C., Minadakis, G., & Papadopoulos, G. A. (2016). Foreshocks and short
term
hazard
assessment of large earthquakes using complex networks: the case of the 2009 L’Aquila earthquake
.
Nonlinear
Processes Geophys, 23(4), 241256.
[
6] Fleming, K., Parolai, S., GarciaAristizabal, A., Tyagunov, S., Vorogushyn, S., Kreibich, H., & Mahlke, H. (2016)
.
Harmonizing
and comparing singletype natural hazard risk estimations. Annals of Geophysics, 59(2), 0216.
[
7] Gill, J. C., & Malamud, B. D. (2016). Hazard interactions and interaction networks (cascades) within multi

hazard
methodologies. Earth System Dynamics, 7(3), 659.
[
8] Van Verseveld, H. C. W., Van Dongeren, A. R., Plant, N. G., Jäger, W. S., & Den Heijer, C. (2015).
Modelling
multi
hazard hurricane damages on an urbanized coast with a Bayesian Network approach. Coastal
Engineering,
103
, 114.
[
9] Rahmati, O., Yousefi, S., Kalantari, Z., Uuemaa, E., Teimurian, T., Keesstra, S., ... & Tien Bui, D. (2019). Multi

hazard
exposure mapping using machine learning techniques: A case study from Iran. Remote Sensing, 11(16
),
1943
.
[
10] Reed, D. A., Friedland, C. J., Wang, S., & Massarra, C. C. (2016). Multihazard systemlevel logit
fragility
functions
. Engineering Structures, 122,1423.
[
11] Lacasa, L., Nicosia, V., & Latora, V. (2015). Network structure of multivariate time series. Scientific reports, 5
,
15508
.
[
12] Aven, T. (2010). On the need for restricting the probabilistic analysis in risk assessments to variability.
Risk
Analysis
:An International Journal, 30(3), 354360.
[
13] Kaplan, S., & Garrick, B. J. (1981). On the quantitative definition of risk. Risk analysis, 1(1), 1127.
[
14] Dubois, D. (2010). Representation, propagation, and decision issues in risk analysis under
incomplete
probabilistic
information. Risk Analysis: An International Journal, 30(3), 361368.
[
15] Aven, T. (2015). The concept of antifragility and its implications for the practice of risk analysis. Risk
analysis,
35
(3), 476483.
[
16] Darabi, H., Choubin, B., Rahmati, O., Haghighi, A. T., Pradhan, B., & Kløve, B. (2019). Urban flood risk
mapping
using
the GARP and QUEST models: A comparative study of machine learning techniques. Journal of
hydrology,
569
,142154.
[
17] Pilkington, S. F., & Mahmoud, H. N. (2016). Using artificial neural networks to forecast economic impact
of
multi
hazard hurricanebased events. Sustainable and Resilient Infrastructure, 1(12), 6383.
[
18] Hamadeh, N., Karouni, A., Daya, B., & Chauvet, P. (2017). Using correlative data analysis to develop
weather
index
that estimates the risk of forest fires in Lebanon & Mediterranean: Assessment versus
prevalent
meteorological
indices. Case Studies in Fire Safety, 7, 822.