PhD Modelling Week, September 4-11, 2016 - Verona (Italy)

European Cooperation in Science and
Technology   EU   Mathematics for Industry Network   Società Italiana di Matemtica Applicata ed
Industriale   Association for Mathematics Applied to Economics and Social
Sciences
Sportello Matmatico per l'Industria Italiana   Department of Computer Science, Univ. of Verona   Department of Mathematics, Univ. of Trento   Camera di Commercio Verona   Confindustria Verona


Home
Projects
Schedule
Venue and accomodation
Registration and fees payment
Participants
Organising Committee and Information
Sponsors

Home

Between September 4 and 11 2016, the joint Doctoral School of Mathematics of the Universities of Trento and Verona and the Department of Computer Science of the University of Verona organise a Modelling Week for PhD students. This event will take place at the Department of Computer Science and will gather PhD students in Mathematical Sciences from different European countries, together with academic supervisors and representatives from industry. This initiative was supported by COST action TD1409, Mathematics for Industry Network (MI-NET). Students work together in groups on projects defined and supervised jointly by academic and industrial cooperation partners, and thus find out how mathematics contributes to the solution of currently relevant industrial problems.

A number of problems from various application contexts will be presented to participants and teams will be formed in order to tackle each problem during the week, under the supervision of an instructor. The goal is to understand the problem sufficiently well in order to select the appropriate modelling and the possible approaches. Then the students will devise and numerically implement a solution, using GNU OCTAVE, MATLAB, C, C++ or other languages or packages. On Saturday, the various teams will present their results to all the participants, inclusive industrial representatives. A final report on the assigned problem will be prepared by each group, within two months.

Projects

1) Simulation of Particle Accelerator Cavities (instructor: Dr. Carlo De Falco, Milan Polytechnic)

Particle accelerators, such as the LHC at CERN, take fundamental particles up to the speed lightby exposing them to properly excited electromagnetic fields. Radio frequency cavities are responsible for the acceleration of the beam. The fields created within these devices are extremely sensitive to the geometry of the cavity itself. A company in Schio (Vicenza, Italy), for which the instructor does a specific consulting, constructs such particle accelerator cavities and needs highly reliable simulations of those fields in function of the geometry of the cavity to customize their products for scientific laboratories. The problem is then, assigned different cavity geometries of interest, to appropriately represent the computational domains through Computer Aided Design (CAD) in order to run simulations of such systems. The instructor will suggest to the team to use isogeometric analysis tools, that allows the exact treatment of geometries defined by CAD even at the coarsest level of discretization refinement. The instructor will provide suitable teaching material to give an introductory overview of CAD geometries and Isogeometric Analysis through the illustration of basic examples, then the team will go through the application of those tools and concepts to the simulation of accelerator cavities having a prescribed geometry.

2) Stress models and optimization models for the banking/industry  sector. Instructors: Dr. Luca Di Persio, University of Verona, in collaboration with IASON Ltd (Dr. Michele Bonollo), and in collaboration with Dr. Ben Paoletti

Banks must develop stochastic models and efficient numerical schemes within the Basel framework, in order to cover the whole risk measurement process, from the parameter estimation to the risk metrics computation. To this extent, a first challenging problem is to design the so called satellite models, i.e. macroeconomic models that, starting from some relevant variables, the so called risk drivers (oil price, GDP, inflation rate, etc..) can propagate the high level regulatory stress test to the effective bank exposures, in order to estimate the granular risk factors (default probability increase of the obligors, hair cut on the real estates collateral values and so on). In this field the reverse stress test is the most challenging problem, i.e. to find which shocks in the high level variables can imply a given output in the internal target bank variables (revenues, provisioning, capital). In other words one has to choose the set of (possibly very different) "input scenarios" that minimize the distance from the target output. More generally, banks and industries very often face some optimization problems, that can vary from the yield management to the optimal scheduling until the risk measures optimization and several other fields. Practical real world problems will be described and assigned.

3) The fundamental review of the trading book (instructors: Dr. Luca Di Persio, University of Verona, in collaboration with Dr. Michele Bonollo (IASON Ltd).

This new framework for the market risk asks to the banks to move from the classical VaR risk measure to the Expected Shortfall. In his new environment the backtesting of the whole model becomes more and more difficult, both for theoretical and computational problems. The applications to a concrete situation will be performed with the collaboration of IASON Ltd.

4) Optimization in city-logistics (supervisor: Prof. Renata Mansini, University of Brescia, Italy, instructors: Dr. Daniele Manerba and Dr. Eng. Marco Colombi)

City logistics deals with the sustainability problems encountered in urban freight transportation. In this context, we will analyze two main types of problems arising in both private and public sectors, suggested by the municipality of Brescia (we aim to involve also the municipality of Verona in those problems). The first problem is a complex procurement and routing problem where a company uses its private fleet of vehicles to visit suppliers offering various products at different prices and quantities. The company aims to select a subset of suppliers so to satisfy products demand at the minimum traveling and purchasing costs. Additional constraints as limited capacity for the vehicles and possible incompatibility among products loaded on the same vehicle are also taken into account. The second problem concerns the public sector and refers to the problem of separate waste collection. In Italy most of the municipalities perform garbage collection in a door-to-door way. Since the elimination of the use of dumpsters, all residents and companies have been provided with a bundle of bins and bags through which to get out the garbage. Nowadays each operator in charge with garbage collection has a predefined zone to serve each day for a specific type of waste, but can freely choose the order to cross all the streets that require a service. We want to show how the costs for the municipalities can be strongly reduced by optimizing the tours for garbage collection. The goal is to show how both problems can be tackled formulating mathematical problems solved by means of commercial mixed integer programming solver (CPLEX) or developing more complex but also more efficient solution algorithms (branch-and-cut and branch-and-price methods). Whenever possible we will also show to the participants how they can obtain traveling times necessary to solve the problems directly from real networks by using open data.

5) Simulation of fluid-dynamic and solute delivery in blood microvessel networks (instructor: Dr. Paola Causin, University of Milano)

Microcirculatory systems comprise thousands of vessels, organized into arterioles (small arteries), capillaries, and venules (small veins), with diameters ranging from a few hundred microns down to a few microns. At present, techniques like positron emission tomography, magnetic resonance imaging and contrast echocardiography allow to study non-invasively regional blood flow in internal organs of human patients. However, the information content of such measurements is not complete, both in baseline conditions as well as in altered conditions and data (geometry, fluid-dynamics, physiological parameters, provocation studies) are often not coherently collected. The instructor will suggest to the team to use a mathematical approach to gain insight into these problems. Computational models will be adopted to solve the fluid-dynamic field in different microvascular networks, both synthetically generated or derived from medical imaging. Several conditions, physiological and pathological will be simulated and compared to available data. Solute (oxygen, carbon dioxide) transport and delivery to the interstitial tissue will be added as well to the model if time allows.

Schedule

The always updated schedule is available here.

Venue and accomodation

The event will take place at the Department of Computer Science, located in the southern campus of the University of Verona. Address: Strada le Grazie 15, 37134 Verona. Concerning accommodation, a certain number of places in shared double room with kitchen facilities at the nearby guesthouse are reserved for people needing a financial support. There are several lodging possibilities in Verona, among others: that have special fares for people visiting the Deparment of Computer Science.

Registration and fees payment

The deadline for the registration is May 31, 2016.
If you are going to request financial support for lodging, travel expenses and/or fee reimbursement, please register within April 30, 2016.
The registration fee is 150.00 Euros (it includes, in particular, coffee breaks, lunches, dinners and final social dinner).
In order to confirm the registration, the fee is due within 15 days from submitting the registration form.

For registration, request of financial support and payment instructions, please go to the registration page.

Participants

Francesco CORDONI,U. of Trento
Chiara BENAZZOLI,U. of Trento
Alessandro MELLA,U. of Verona
Matteo FRIGO,U. of Verona
Sanja RUZICIC,U. of Novi Sad
Gennaro CIBELLI,U. of Modena
Chiara PIAZZOLA,U. of Innsbruck
Carlo COMIN,U. of Trento
Marcella BONAZZOLI,U. of Nice
Christian MUENCH,TU Munich
Martina PRUGGER,U. of Innsbruck
Roberto ZANOTTI,U. of Brescia
Simon RABANSER,U. of Innsbruck
Mauro BONAFINI,U. of Trento
Elena GABURRO,U. of Trento
Giovanni MARINI,U. of Trento
Federico REALI,U. of Trento
Chiara VENTURINI,U. of Verona
Marcus CORDI,Centrale Supelec
Massimiliano ROSSI,U. of Verona
Andrea CRACCO,U. of Verona
Emanuele CITTADINO,U. of Verona
Filippo MELA,U. of Verona
Ihor SHARMANSKY,U. of Verona
Renata MANSINI,U. of Brescia
Daniele MANERBA,U. of Brescia
Beniamino PAOLETTI,Iason Roma
Michele BONOLLO,IMT Lucca
Paola CAUSIN,U. of Milan
Carlo DEFALCO,Polytechnic Milan
Luca DIPERSIO,U. of Verona
Romeo RIZZI,U. of Verona
Marco CALIARI,U. of Verona
Antonio MARIGONDA,U. of Verona
Giandomenico ORLANDI,U. of Verona
Giacomina BRUTTOMESSO,U. of Verona
Aurora MIORELLI,U. of Verona

Organising Committee and Information

TitleNameSurnameJob TitleRole
Dr.LucaDi PersioResearcher, U. of Veronaproblem selection/supervision
Prof.RomeoRizziAssoc. Professor, U. of Veronainstructor
Prof.RenataMansiniAssoc. Professor, U. of Besciaproblem selection/supervision
Prof.MarcoCaliariAssoc. Professor, U. of Veronaweb site/dissemination
Dr.MicheleBonolloConsultant, IASON LTDinstructor
Dr.AntonioMarigondaResearcher, U. of Veronaproblem selection/supervision
Prof.GiandomenicoOrlandiProfessor, U. of Veronacoordinator

For more information, please write an email to master.math@ateneo.univr.it.

Sponsors