[Forum SIS] Avviso di seminario: Tarantola@ Dipartimento di Statististica e Metodi Quantitativi Milano-Bicocca

[Fulvia Pennoni]

********************************************************
Dipartimento di Statistica e Metodi Quantitativi
Via Bicocca degli Arcimboldi, 8 - 20126 Milano
*********************************************************

"THE ROLE OF GLOBAL SENSITIVITY ANALYSIS IN SIMULATION MODELLING"

Ing. Stefano Tarantola
European Commission, Joint Research Centre

Mercoledì  29 Gennaio ore 15.00 aula seminari 4026 Ed. U7  IV piano
__________________________________________________________

Abstract:
Mathematical modellers and regulatory agencies worldwide share the 
belief that sensitivity analysis
is a key ingredient of the quality of a model-based study. According to 
the European Commission
"/[...] sensitivity analysis can be used to explore how the impacts of 
the options you are analyzing//
//would change in response to variations in key parameters and how they 
interact"/. The term interaction
is also used in the guidelines for modellers of the US environmental 
Protection Agency:
"/Sensitivity//
//analysis methods should preferable be able to deal with a model 
regardless to assumptions//
//about a model's linearity and additivity, consider interaction effects 
among input uncertainties [...]"/,
and evaluate the effect of an input while all the other inputs are 
allowed to vary as well.
In spite of this call for the use of global tools for sensitivity 
analysis, plenty of cases are found in the
literature where local sensitivity analysis is employed. In local 
analyses factors' importance is investigated
by derivative of the output with respect to that input; such derivatives 
are informative only at
the base point where they are computed, but do not provide for an 
exploration of the rest of the
space of the input factors unless some conditions are met in the 
mathematical formulation under
analysis. In this talk I will illustrate the properties of global 
sensitivity analysis approaches and will
discuss variance-based methods, the best practice available today.


Fulvia Pennoni

-------------- parte successiva --------------
Un allegato HTML è stato rimosso...
URL: <http://www.stat.unipg.it/pipermail/sis/attachments/20140120/f838a32b/attachment.html>