- NumPI - Mailing list — Dipartimento di Matematica, UniPI

Annuncio di seminario
by Leonardo Robol 12 Dec '18

12 Dec '18

Buongiorno, allego l'annuncio del seminario per la settimana prossima, che sarà tenuto da Giulio Masetti (ISTI-CNR). Colgo l'occasione per ricordarvi l'appuntamento di domani, per il seminario di Stefano Massei, alle 10:00 in Aula Seminari Ovest, Informatica. -- Leonardo Robol. ==== Speaker: Giulio Masetti Affiliation: ISTI-CNR Time: Thursday, 13 December 2018, h. 10:00 Place: Aula Seminari, Dipartimento di Matematica Title: Computing dependability-oriented measures on Markov chains by means of matrix functions In the context of computing and communication system, dependability is defined as the ability to deliver service that can justifiably be trusted. Dependability is then an umbrella term that encompasses several attributes, such as: availability, reliability, safety, integrity, maintainability, and confidentiality. Design and analysis of models capable to express functional specifications and behavioural aspects is an important part of the system dependability justification process. In particular, Markov chains reflect important characteristics of computing systems such as discrete states and memoryless property, and then are often employed to model complex systems. Starting from the model Markovian process, other processes can be defined to express gains or losses according to which state the Markov process is in, and then all the mentioned dependability attributes can be evaluated as measures on these processes. More and more complex systems are designed, and then new techniques are required to tackle the evaluation of dependability-oriented measures on large models. In a world where systems comprise hundreds or thousands of interconnected components, justifying dependability is an ever increasing challenge. The main contribution presented in this talk is, focusing on Continuous Time Markov Chains, recasting dependability-oriented measures as the evaluation of a bilinear form where the matrix is indeed a matrix function of the infinitesimal generator matrix characterizing the Markovian stochastic process. In particular, chains with absorbing states, relevant to evaluate system reliability and connected attributes, represent one of the main challenges the modeling community has to deal with, and -- we think -- one of the cases where applying results from the matrix functions body of knowledge can have the greater impact.

1 1

Annuncio di seminario
by Leonardo Robol 29 Nov '18

29 Nov '18

Buongiorno, allego l'annuncio del seminario per la settimana prossima, che sarà tenuto da Stefano Massei (EPF Lausanne). L'appuntamento è giovedì 6 dicembre alle 10:00 presso il Dipartimento di Informatica, Sala Seminari Ovest. -- Leonardo Robol. ==== Speaker: Stefano Massei Affiliation: EPF Lausanne Time: Thursday, 6 December 2018, h. 10:00 Place: Sala Seminari Ovest, Dipartimento di Informatica. Title: A low-rank technique for computing the quasi stationary distribution of Galton-Watson processes Branching processes describe the dynamics of a population of individuals which reproduce and die independently, according to some specific probability distributions. More precisely, we assume that any individual has a unit lifetime, at the end of which it might give birth to one or more offsprings simultaneously. This is encoded into the probability generating function P(z):=\sum p_j z^j where p_j is the probability of generating j individuals. These kind of processes are known in the literature as Galton-Watson processes. We consider populations that are certain to become extinct, yet appear to be stationary over any reasonable time scale. More precisely, we are interested in characterizing the quasi-stationary distribution of the process, i.e., the asymptotic distribution of the population size, conditional on its survival. Yaglom proved that if m:=P'(1)<1 then the quasi stationary distribution exists and its probability generating function G(z):=\sum g_j z^j solves the Schroeder functional equation G(P(z))= mG(z)+1-m, z in [0,1]. (*) We study the link between the regularity of P(z) and that of G(z) and we propose a strategy for solving (*) in the case where P(z) and G(z) are analytic on a disc of radius r>1. We see that the discretization of (*) leads to a numerical method that is capable to find arbitrary accurate approximations of the coefficients of G(z). Moreover, we point out the (numerical) low-rank structure that appears in the discretized problem, and we show how to exploit it in the proposed procedure.

1 0

Annuncio di seminario
by Leonardo Robol 26 Nov '18

26 Nov '18

Buongiorno, allego l'annuncio del seminario di Marcus Carlsson, programmato per la prossima settimana. Il seminario si terrà in SNS, e sarà di martedì e non di giovedì. -- Leonardo Robol. ==== Speaker: Marcus Carlsson Affiliation: Lund University Time: Tuesday, 27 November 2018, h. 11:00 Place: Aula Bianchi scienze, SNS Title: Perturbation theory for the spectral decomposition of matrices Consider a Hermitian matrix and a small perturbation. We will consider how the eigenvalues and eigenvectors change with the perturbation. Typical theorems concerning this either consider a perturbation along a line or provide general estimates valid for all perturbations (such as Weyl’s inequalities). We first give an overview of the field and then provide new results which are global in the sense that the perturbation is not restricted to a line, but local in the sense that the estimates are valid modulo certain ordo terms. We apply this to the functional calculus of the matrix square root and matrix modulus. In the kernel-free case the corresponding result follows from the so called Daleskii-Krein theorem.

1 1

Avviso di seminario (e reminder)
by Leonardo Robol 22 Nov '18

22 Nov '18

Buongiorno, allego l'annuncio del secondo seminario nella serie NumPI, e con l'occasione vi ricordo l'appuntamento di domani alle 10:00 per quello di Giovanni Barbarino (Sala Seminari Est, Dipartimento di Informatica). ==== Speaker: Jie Meng Affiliation: Pusan National University Time: Thursday, 22 November 2018, h. 10:00 Place: Sala Seminari, Dipartimento di Matematica, Università di Pisa Title: On the positive definite solution of the nonlinear matrix equation X^p=A+M^T(X#B)M The nonlinear matrix equation X^p = A + M^T (X # B)M is considered, where p ≥ 1 is a positive integer, M is an n × n nonsingular matrix, A is a positive semidefinite matrix and B is a positive definite matrix. We call C # D the geometric mean of positive definite matrices C and D. Some properties of the positive definite solution of the nonlinear matrix equation is investigated, including the existence and uniqueness of the solution, a lower and an upper bound. Iteration method for finding the numerical solution is proposed. Perturbation bounds with respect to small perturbations on the coefficient matrices are obtained. -- Leonardo Robol

1 1

Fwd: [dynlist] Dyn.Sys - C. Wormell @ Centro De Giorgi 14, 00h - 14th Nov.
by Federico Poloni 13 Nov '18

13 Nov '18

Mi hanno segnalato questo seminario domani al Centro De Giorgi --- non è uno della nostra serie, ma è comunque un argomento che può essere interessante per alcuni. Cordialmente, >> Begin forwarded message: >> >> *From: *Paolo Giulietti <paologiulietti.math(a)gmail.com >> <mailto:paologiulietti.math@gmail.com>> >> *Subject: **[dynlist] Dyn.Sys - C. Wormell @ Centro De Giorgi 14,00h - >> 14th Nov.* >> *Date: *November 12, 2018 at 5:35:20 PM GMT+1 >> *To: *dynlist-pisa(a)googlegroups.com <mailto:dynlist-pisa@googlegroups.com> >> >> Dear Dynamical Friends, >> >> this week, on Wednesday 14th of November at 14:00h at the Centro de >> Giorgi, ground floor, C. Wormell from the University of Sydney will >> talk about "Chebyshev Galerkin methods for transfer operators in >> uniformly-expanding dynamics" (abstract below). >> >> Let me Remind you that at 15:00h at Palazzo della Carovana there will >> be the monthly science colloquium of the SNS. >> >> Moreover, if you are using the GoogleCalendar, check it out to see >> latest additions to the January/February schedule! >> >> Cheers! >> P&J >> >> ----------------------------------------------------------------------------------------------------------------------------------- >> >> Abstract: Full-branched expanding maps, a class of simple chaotic >> systems, are >> commonly used as models for chaotic dynamics, as well as being of >> number-theoretical interest. However, nonspecialised numerical methods >> to calculate long-time statistical quantities such as invariant measures >> have a poor trade-off between computational effort and accuracy. >> >> In this talk I will present a rigorous method to calculate statistics of >> these maps by discretising the transfer operator in a Chebyshev >> polynomial basis. This is a highly efficient discretisation: I will show >> that, for analytic maps, numerical estimates obtained with this method >> converge exponentially quickly in the order of the discretisation. >> Finally, I will discuss some implementations of these algorithms for >> exploration and rigorous proof: in particular, in only a fraction of a >> second on a personal computer one can produce estimates of most >> statistical properties accurate to 14 decimal places. >> >> >> -- >> ------------------------------------------------------------------------- >> "NO! Try not. Do, or do not. There is no try." >> >> "Much to learn, you still have." >> >> \\Yoda\\ <smb://Yoda//> >> >> -- >> Hai ricevuto questo messaggio perché sei iscritto al gruppo >> "dynlist-pisa" di Google Gruppi. >> Per annullare l'iscrizione a questo gruppo e non ricevere più le sue >> email, invia un'email a dynlist-pisa+unsubscribe(a)googlegroups.com >> <mailto:dynlist-pisa+unsubscribe@googlegroups.com>. >> Per postare in questo gruppo, invia un'email a >> dynlist-pisa(a)googlegroups.com <mailto:dynlist-pisa@googlegroups.com>. >> Visita questo gruppo all'indirizzo >> https://groups.google.com/group/dynlist-pisa. >> Per altre opzioni visita https://groups.google.com/d/optout. > -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Fwd: 2 open PhD positions, TU Eindhoven with industry in Milano
by Gianna M. Del Corso 06 Nov '18

06 Nov '18

Michiel Hochstenbach mi chiede di diffondere questa call per 2 posizioni di PhD a Eindhoven/Milano. Gianna > Inizio messaggio inoltrato: > > Da: "Hochstenbach, M.E." <M.E.Hochstenbach(a)tue.nl> > Oggetto: 2 open PhD positions, TU Eindhoven with industry in Milano > Data: 6 novembre 2018 15:10:43 CET > A: "delcorso(a)di.unipi.it" <delcorso(a)di.unipi.it> > > Dear Gianna, > > I hope you are fine! > > This email is to let you know that we have 2 open 4-year PhD positions at TU Eindhoven per March 1, 2019 in the "BIGMATH" project; I will be the main academic advisor. > These are 2 Marie Curie EID (European Industrial Doctorate) positions, together with industry in Milano (AcomeA and SDG). > The PhD students will spend ca 28 months in Eindhoven and 20 months in Milano. > > Your friend Alessandra Micheletti is also involved in this project. > > All information is available at http://www.win.tue.nl/~hochsten/bigmath.html. > The project descriptions can be found there; the topics might include numerical linear algebra, big data, large-scale optimization, modeling, probability and statistics, model reduction. > The deadline is December 5. > > Could you please forward this info to potentially interested students? > > Thank you very much in advance! All the very best and kind regards, > > Michiel Hochstenbach > > PS: There are also 2 positions at the University of Milano (please see http://itn-bigmath.unimi.it/); for those positions the restriction is a presence of <= 12 months in Italy over the last 3 years (this is not a restriction for the Eindhoven positions). Moreover, there are 2 positions in Lisbon and 1 in Novi Sad. >

1 0

Avviso di seminario
by Leonardo Robol 01 Nov '18

01 Nov '18

Speaker: Giovanni Barbarino Affiliation: Scuola Normale Superiore, Pisa Time: Thursday, 15 November 2018, h. 10:00 Place: Sala Seminari Est, Dipartimento di Informatica, Università di Pisa Title: Higher order spectral symbols and eigenvalues approximation for Toeplitz matrices When modeling natural phenomena with linear partial differential equations, the discretized system of equations is generally represented by a sequence of matrices, often coupled with a function called spectral symbol that enclosures asymptotic spectral information of the sequence. In fact, when sampling the symbol, we obtain an approximation of the spectrum of the matrices, with an error having magnitude inversely proportional with respect to the matrix size. Under particular hypothesis, the sequence presents an higher-order asymptotic formula for the eigenvalues, that let us precisely compute the eigenvalues of large matrices with a technique of interpolation-extrapolation at a low computational cost and that is also parallelizable. This algorithm seems also to work on different kind of matrices, but with irregularities on the extreme eigenvalues, for which a more accurate analysis is needed. -- Leonardo Robol

1 0

Seminari di Analisi Numerica
by Leonardo Robol 09 Oct '18

09 Oct '18

Carissimi, come negli scorsi anni ci piacerebbe organizzare una serie di seminari di analisi numerica, con lo scopo di conoscere le attività all'interno del gruppo e dare spazio a studenti e visitatori. Per cercare di massimizzare la possibilità di seguirli, chiederei a tutti gli interessati di compilare il Doodle [1], segnalando gli slot in cui siete liberi da lezioni e/o che preferite. Poi cercheremo di tenerne conto nel programmare i seminari. Ultimo, ma non meno importante, siamo in cerca di volontari per tenere qualche seminario! Per proporvi, o per proporre visitatori e/o studenti scrivete una e-mail a me e Federico, eventualmente con la disponibilità di date. Può essere una buona occasione per provare una presentazione prima di una conferenza o di una tesi, per esempio. Aspettiamo con fiducia adesioni numerose! [1]: https://doodle.com/poll/eqfkkrztkzywp2qp Ciao, -- Leonardo e Federico.

1 0

corso dottorato @SNS "Metodi di Krylov e applicazioni"
by Federico Poloni 25 Sep '18

25 Sep '18

Buongiorno, segnalo, per gli interessati, questo corso di dottorato che si terrà alla Scuola Normale Superiore. Le lezioni (per circa 60 ore) si terranno a partire dal 5 novembre col seguente orario: Lunedi ore 11-13 Aula Bianchi (Scienze) Mercoledi ore 11-13 Aula Tonelli presso il Palazzo della Carovana in Piazza dei Cavalieri. Indicativamente le lezioni termineranno a fine maggio. Ulteriori dettagli nella scheda allegata. Cordialmente, -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Fwd: Master in High Performance Computing, Trieste
by Federico Poloni 06 Jul '18

06 Jul '18

Ricevo e pubblicizzo questo annuncio di master. Un saluto a tutti. =============================================================================== Buongiorno, avrete piacere di sapere e di condividere che le iscrizioni <https://www.mhpc.it/how-apply> per la 5a edizione del Master in High Performance Computing <https://www.mhpc.it/> (MHPC) sono state estese *_fino al 24 luglio_*. Al momento abbiamo 5 borse sponsorizzate da beanTech <www.beantech.it>, Plus Solutions <https://www.plusolutions.it/>, OGS <www.ogs.trieste.it/en> e INFN. <www.ts.infn.it/en> Si tratta di un percorso professionalizzante post-laurea (verranno considerati solo candidati in possesso di una laurea magistrale o dottorandi/con dottorato di ricerca) e si svolge a Trieste in un ambiente di ricerca internazionale e stimolante, costituito da centri di eccellenza come SISSA <https://www.sissa.it/> e ICTP <https://www.ictp.it/>. *Obiettivo del master è preparare in 12 mesi specialisti per l'accademia e l'industria nel settore dell'HPC/Big Data, in cui le offerte di lavoro sono in crescita costante. *** Nelle prime 3 edizioni del master sono stati formati 33 specialisti, il 100% di loro ha trovato un'occupazione nel proprio settore di competenza, sia in ambito accademico che industriale, ed un terzo di loro ha già una posizione a tempo indeterminato. Le lezioni sono tenute in lingua inglese da docenti SISSA e ICTP e da esperti di fama internazionale, provenienti da realtà come Intel e Nvidia. Sperando di aver suscitato il vostro interesse verso la nostra iniziativa, vi preghiamo di far girare questa comunicazione grazie ai vostri canali. Vi invitiamo a seguirci al Campus Party di Milano <italia.campus-party.org> il 20 Luglio alle ore 17 (sezione Coding Stage). Rimaniamo a vostra disposizione <mailto:mhpc@sissa.it> per ulteriori informazioni. Cordialmente, lo staff di MHPC Twitter <https://twitter.com/mhpc_sissa_ictp> Facebook <https://www.facebook.com/mhpc.sissa.ictp/> LinkedIn <https://www.linkedin.com/company/mhpc---sissa---ictp/> =============================================================================== Dear Sir/Madame, you will be interested to know and to share that the enrollment <https://www.mhpc.it/how-apply> for the 5th edition of the Master in High Performance Computing <https://www.mhpc.it/> (MHPC) has been extended to a new *_deadline: July 24th_*. We currently offer 5 full scholarships sponsored by beanTech <www.beantech.it>, Plus Solutions <https://www.plusolutions.it/>, OGS <www.ogs.trieste.it/en> and INFN. <www.ts.infn.it/en> This master is a post-graduation school that is organized in Trieste by the synergistic cooperation of two excellence research centers: SISSA <https://www.sissa.it/> and ICTP <https://www.ictp.it/>. *Our aim is to prepare, within 12 months of master, HPC/Big Data specialists, which is a fast growing job in either academy and industry settings. *** During the first three editions of this master, 33 people have been educated, 100% of them are working in their area of expertise and one third of them have already a permanent job. The course is kept in english by SISSA/ICTP lecturers and world-renowed experts from both academy and industry, such as Intel and Nvidia. Hoping that we interested you with this short mail, we would be grateful if you could forward this communication to your networks. You can meet us at the Milano Campus Party <italia.campus-party.org> July 20th 5pm (Coding Stage section). For any further information, please write to us <mailto:mhpc@sissa.it>. Yours faithfully, MHPC staff Twitter <https://twitter.com/mhpc_sissa_ictp> Facebook <https://www.facebook.com/mhpc.sissa.ictp/> LinkedIn <https://www.linkedin.com/company/mhpc---sissa---ictp/>

1 0

Pisa-Hokkaido-Tor Vergata Summer School in Mathematics, 27 Agosto-8 Settembre
by Federico Poloni 28 Jun '18

28 Jun '18

Buongiorno, segnalo questa scuola per i possibili interessati. Si tratta di una scuola dedicata a studenti di master/dottorato di matematica, con diversi mini-corsi di 4-8 ore su vari argomenti (tra cui alcuni più numerici/applicati). ``` PISA-HOKKAIDO-ROMA2 Summer School on Mathematics and Its Applications 2018 Centro De Giorgi, Pisa, 27 Agosto - 8 Settembre 2018. Nell'ambito delle attività congiunte dei Dipartimenti di Matematica della Università di Pisa, di Roma Tor Vergata e di Hokkaido University, dal 27 Agosto all'8 Settembre 2018, al Centro De Giorgi, a Pisa, si terrà la prossima scuola estiva. I corsi offerti riguardano diverse aree della matematica pura e applicata: Paola Boito and Federico Poloni: "Matrix functions and network analysis Claudio Bonanno: "Infinite Ergodic Theory" Piero D'Ancona: "Dispersive estimates and application" Martina Lanini: "Introduction to Schubert varieties" Ugo Locatelli: " The KAM theorem, following the original approach designed by Kolmogorov" Jun Masamune: " Global Analysis - the heat kernel, old and new" Maurizio Parton: "Introduction to neural networks for mathematicians" Simona Settepanella " Discrete Morse Theory" A questo indirizzo si trovano tutte le indicazioni e le informazioni a riguardo: http://www.crm.sns.it/event/432/ In particolare: 1) La scuola è gratuita. 2) Per partecipare alla scuola occorre presentare domanda, seguendo le istruzioni sul sito. 3) L'università può erogare crediti per la frequenza ed il superamento degli esami della scuola. Per l'erogazione dei crediti è necessaria una iscrizione ulteriore e più formale, le cui modalità saranno comunicate agli interessati. 4) All'indirizzo http://www.crm.sns.it/event/432/speakers.html?page=1#title si trovano gli abstract dei corsi. ``` -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Seminario Veneziani
by Beatrice Meini 20 Jun '18

20 Jun '18

Buongiorno, vi inoltro questo avviso di seminario al dipartimento di Matematica, che potrebbe essere di interesse per il gruppo. Beatrice ---- Giovedi 28 giugno, ore 11:30, Aula Magna, Dipartimento di Matematica Prof. Alessandro Veneziani, Emory Center for Mathematics and Computing in Medicine, Atlanta Title:Old and new challenges of incompressible computational fluid dynamics in cardiovascular problems Abstract: With the progressive use of computational fluid dynamics for supporting medical doctors decision in clinical treatment of cardiovascular diseases, new and old challenges arise in terms of efficiency and reliability. In this talk, we will consider some of them, particularly related to (1) the efficient stabilization and solution of incompressible steady flows (as a surrogate of time average solutions over a heartbeat); (2) the reliable treatment of defective boundary data problems. Applications will be illustrated to coronary as well as aortic diseases.

1 0

seminario Iliescu
by Federico Poloni 14 Jun '18

14 Jun '18

Salve, ricevo e inoltro questo avviso di seminario che potrebbe essere di interesse per gli iscritti a questa lista. === Avviso di seminario: Martedi 19 giugno 2018, ore 1500 presso il dipartimento di Matematica Aula Magna Traian Iliescu (Virginia Tech) Title: "Physically-Constrained Data-Driven Correction for Reduced Order Modeling of Fluid Flows" Abstract: We propose a data-driven correction reduced order model (DDC-ROM) framework for the numerical simulation of fluid flows, which can be formally written as DDC-ROM = Galerkin-ROM + Correction. The new DDC-ROM is constructed by using ROM spatial filtering and data-driven ROM closure modeling (for the Correction term). Furthermore, we propose a physically-constrained DDC-ROM (CDDC-ROM), which aims at improving the physical accuracy of the DDC-ROM. The new physical constraints require that the CDDC-ROM operators satisfy the same type of physical laws (i.e., the Correction term's linear component should be dissipative and the Correction term's nonlinear component should conserve energy) as those satisfied by the fluid flow equations. To implement these physical constraints, in the data-driven modeling step, we replace the unconstrained least squares problem with a constrained least squares problem. We perform a numerical investigation of the new CDDC-ROM and standard DDC-ROM for a 2D channel flow past a circular cylinder at Reynolds numbers $Re=100, Re=500$, and $Re=1000$. To this end, we consider a reproductive regime as well as a predictive (i.e., cross-validation) regime in which we use as little as $50\%$ of the original training data. The numerical investigation clearly shows that the new CDDC-ROM is significantly more accurate than the DDC-ROM in both regimes. -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

mincorso su "Functions of matrices"
by Federico Poloni 26 May '18

26 May '18

Buongiorno! Segnalo questo mini-corso di 12 ore che si terrà in SNS nelle prossime due settimane. "Functions of matrices: theory, numerical methods and applications" Prof. Michele Benzi (Emory University, Atlanta, USA) La prima lezione sarà lunedì in aula Fermi (Scuola Normale) alle 11:00. In caso ci fosse qualcuno extra-SNS interessato a partecipare vi prego di spedirmi un'e-mail così tengo traccia dei numeri di partecipanti esterni. Arrivederci, -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Seminario G. Barbarino - poll
by Federico Poloni 21 Apr '18

21 Apr '18

Buongiorno! Vorremmo far tenere un seminario a Giovanni Barbarino nelle prossime settimane, sul tema "Higher Order Spectral Symbols and Eigenvalues Approximation for Toeplitz Matrices", e volevo chiedervi scegliere tra le varie possibilità per l'orario del seminario. Se siete interessati a seguire il seminario, segnate le vostre disponibilità di massima su https://doodle.com/poll/vv96x5zbth5cbss4 e poi sceglierò l'opzione più gettonata. Grazie e a risentirci per la data definitiva. Cordialmente, -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 2

Avviso di seminario
by Dario A. Bini 30 Mar '18

30 Mar '18

Buongiorno, vi invio gli estremi del seminario di Raf Vandebril. Dario --- Giorno: 6 Aprile, ore 10:00 Luogo: Aula Magna, DIpartimento di Matematica Autore: Raf Vandebril, KU Leuven Titolo: On classical, extended, and rational Krylov and the associated QR algorithms Abstract: In this lecture we will discuss three main classes of Krylov and QR type methods: the classical, the extended, and the rational versions. We will initially focus on the classical QR method and reintepret all ingredients such as: associated Krylov spaces, associated Hessenberg matrices, implicit theorems, bulge chasing, eigenvalue swapping, and subspace iteration. Another look at these essentials will allow us to straightforwardly generalize all of this to the extended and rational case. More precisely, we will start by deducing the structure of the recurrence pencil associated to all Krylov methods. It will be shown that the rational Krylov method produces the most general pencil, namely two Hessenberg matrices, which implicitly store the poles of the rational Krylov method. We will revisit the implicit Q-Theorem and link it to its dual theorem, the implicit HK-Theorem, allowing us to provide a pole chasing algorithm for two Hessenberg matrices. This generalizes the classical and extended QR algorithms. Correctness will follow directly from the previous analysis of the classical QR algorithm. The rational QR algorithm allows for additional freedom, which is the introduction of poles. Even without a good pole-strategy we will see that the iteration count reduces with almost 10 procent.

1 0

Fwd: opportunità nella geofisica computazionale - INGV Pisa
by Dario A. Bini 25 Feb '18

25 Feb '18

Mi viene chiesto di diffondere questa informazione. Provvedete anche voi. Dario --- -------- Forwarded Message -------- Subject: opportunità nella geofisica computazionale - INGV Pisa Date: Sat, 24 Feb 2018 10:35:55 -0700 From: Mattia De' Michieli Vitturi <mattia.demichielivitturi(a)ingv.it> To: stefano.galatolo(a)unipi.it, giovanni.federico.gronchi(a)unipi.it, dario.bini(a)unipi.it CC: Tomaso Esposti Ongaro <tomaso.espostiongaro(a)ingv.it> Buongiorno, Vi disturbo perché avremonei prossimi mesi diverse opportunità per assegni di ricerca pre/post-dottorali annuali su tematiche di geofisica computazionale all'INGV di Pisa. Prima di procedere con i bandi, vorremmo esplorare possibili candidature. In particolare, per: 1) data assimilation in atmospheric dispersal models 2) High Performance Computing of multiphase volcanic flows 3) Modeling/simulation of flow and mechanics of porous media 4) Development of automated procedures for the integration of monitoring data into numerical models of volcanic plumes Il profilo 1) sarebbe più matematico applicato, il 2) più computazionale, il 3) più fisico/numerico, il 4) informatico. Se avete in mente persone potenzialmente interessate potete indirizzarle a me o al collega Tomaso Esposti Ongaro per maggiori informazioni: mattia.demichielivitturi(a)ingv.it <mailto:mattia.demichielivitturi@ingv.it> tomaso.espostiongaro(a)ingv.it <mailto:tomaso.espostiongaro@ingv.it> Tutte le attività sarebbero presso l'INGV a Pisa, ma nel caso 2) la borsa è finanziata da OGS (Istituto Naz di Oceanografia e Geofisica Sperimentale) di Trieste. Grazie, buona giornata e buon lavoro Mattia e Tomaso -- Mattia de' Michieli Vitturi, PhD Istituto Nazionale di Geofisica e Vulcanologia Sezione di Pisa Via della Faggiola 32, I-56126 Pisa, Italia ph.: +39-050-8311923 fax.: +39-050-8311942 email: mattia.demichielivitturi(a)ingv.it <mailto:mattia.demichielivitturi@ingv.it> skype: demichie

1 0

Numerical analysis seminars - Durastante
by Federico Poloni 05 Feb '18

05 Feb '18

Buongiorno, inoltro questo avviso di seminario. > il giorno 14 Febbraio alle ore 11 in sala seminari est presso il dipartimento di informatica > il Dott. Fabio Durastante assegnista presso il Dipartimento di Informatica dell'Universita` di Pisa > terra` un seminario dal titolo "Preconditioning Techniques for Some Large Localized or Structured Linear Systems with Application to Discrete Fractional Calculus". Cordialmente, -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Posizione PhD a Strathclyde
by Federico Poloni 20 Dec '17

20 Dec '17

Ricevo e inoltro in caso ci siano persone interessate. > PhD Studentship on novel preconditioned iterative solvers for radial basis function linear systems > > Applications are invited for a PhD studentship on novel preconditioners for radial basis function linear systems, under the supervision of Dr Jennifer Pestana and Dr Alison Ramage, at the Department of Mathematics and Statistics at the University of Strathclyde, Glasgow, Scotland. The project will suit students with interests in numerical analysis and linear algebra. > > The studentship covers UK/EU tuition fees and comes with an annual tax-free stipend at the standard UK rate. International students who can fund the difference between UK/EU and International fee rates are also encouraged to apply. > > The project will begin 1 March, 2018, or earlier by arrangement. > > Further details are available at > > http://personal.strath.ac.uk/jennifer.pestana/RBF_Project.pdf > > Informal enquiries to Jennifer Pestana (jennifer.pestana(a)strath.ac.uk) are welcome. Buone feste a tutti, -federico -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

due giorni a Padova
by Dario Bini 27 Nov '17

27 Nov '17

Carissimi, vi ricordo che l'8 e il 9 febbraio 2018 si terra' la "Due Giorni di Algebra Lineare Numerica e Applicazioni" www.math.unipd.it/~pd2ggaln Se interessati vi invito a comunicare all'indirizzo pd2ggaln(a)math.unipd.it la vostra manifestazione di interesse possibilmente specificando se si desidera tenere una comunicazione. Ciao, Dario ---

1 0

assegno di ricerca UNIPI
by Federico Poloni 21 Nov '17

21 Nov '17

Luca Gemignani mi comunica che è stato pubblicato il seguente bando per un assegno di ricerca sul progetto "Modelli e algoritmi innovativi per problemi strutturati e sparsi di grandi dimensioni" (scadenza domande 20 dicembre; necessario avere il dottorato). Info su https://www.unipi.it/ateneo/bandi/assegni/pra2017/inf/20dic2017/index.htm Un saluto a tutti, -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Seminar
by Dario Bini 02 Nov '17

02 Nov '17

Avviso di seminario: Speaker: Samandar Babaev (Bukhara State University, Uzbekistan) Date: Wednesday, November 8 Time: 11:00 Place: Sala Seminari, Dipartimento di Matematica Title: Optimal Interpolation Formulas in W_2^{(m,m-1)} space Abstract: In this talk, using S.L.Sobolev’s method, optimal interpolation formulas are constructed in /W_{2} ^/(/m,m//−/1)(0/,/1) space. Explicit formulas for coefficients of optimal interpolation formulas are obtained. Some numerical results are presented.

1 0

Avviso di Seminario
by Dario Bini 26 Oct '17

26 Oct '17

Avviso di seminario: Speaker: Federico Poloni (Università di Pisa) Date: Tuesday 2017-10-31 Time: 11:00 Place: Sala Seminari Ovest, Dipartimento di Informatica Title: Counting Fiedler pencils using diagrams Fiedler pencils are a family of matrix pencils which generalizes the well-known companion matrix: they are*linearizations*, i.e., they provide a method to construct, given a matrix polynomial, a linear eigenvalue problem with the same eigenvalues and multiplicities. Fiedler pencils are constructed as products of special block matrices that act nontrivially only on two contiguous blocks. They have a rich structure that give rise to many combinatorial properties. In this talk, we introduce a notation that associates to each pencil a diagram that depicts its action on the blocks. Using this notation, we can obtain visual proofs of several statements in the theory, and we can solve several counting properties (such as "how many distinct Fiedler pencils with repetitions of a given dimension exist"). Among them, in particular, we are interested in counting Fiedler pencils associated to symmetric and palindromic matrix polynomials which preserve the same structure. This talk is based on a joint work with Gianna Del Corso.

1 0

Avviso di seminario
by Dario A. Bini 13 Oct '17

13 Oct '17

Invio il seguente avviso di seminario Dario --- Gorno: Martedi 17 ottobre Ora: 11:00 Aula: Seminari, Dipartimento di Matematica Speaker: Leonardo Robol, ISTI --- CNR Titolo: Backward error analysis for polynomial rootfinders. Abstract: The most common methods to compute roots of polynomials is to rephrase the problem in linear algebra terms: one constructs a companion matrix (or pencil) that has the roots of the polynomial as eigenvalues. However, even when the eigenvalue method used is backward stable, it's non-trivial to map the backward error back on the polynomial. This problem has been studied extensively in the past and it has been shown that relying on the QR method does not provide a normwise backward-stable rootfinder. For this reason, one needs to rely on QZ + appropriate scaling. However, companion matrices are unitary plus rank 1, and we claim that relying on fast QR method that exploits this structure can lead to a backward-stable method. To this end, one needs to ensure that the errors on the unitary and rank 1 parts satisfy different bounds. We show how to perform this analysis of "mixed perturbation backward-error", as well as a method that satisfies these constrains. Time permitting, we discuss some more details that one should consider. We show, for example, that normwise backward stable operations (i.e., unitary rotations in our case) do not necessarily imply backward stable operations on the unitary and low-rank factors, unless some particular identities involving sines and cosines are preserved. This is joint work with Jared Aurentz, Thomas Mach, Raf Vandebril, and David Watkins.

1 0

Optimization & Numerical Analysis seminars: Iannazzo
by Federico Poloni 26 Jul '17

26 Jul '17

Speaker: Bruno Iannazzo Affiliation: Università di Perugia, Italy Time: Thursday 2017-08-27, h. 11:00 Place: Sala Seminari Est, Dipartimento di Informatica, Università di Pisa Title: The Riemannian Barzilai-Borwein algorithm and the matrix geometric mean computation Abstract: We consider the problem of computing the geometric mean of positive definite matrices, the so-called Karcher mean. For this purpose, we adapt the Barzilai-Borwein method from nonlinear optimization to Riemannian manifold optimization. Incomporating a nonmonotone line-search strategy yields global convergence under standard assumptions, fulfilled in the problem of the matrix geometric mean. By a set of numerical tests, we show that the Riemannian Barzilai-Borwein method notably outperforms existing first-order optimization methods for computing the matrix geometric mean. === Everyone is welcome. Sorry for the short notice! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 1

Avviso di Seminario
by Dario A. Bini 11 Jul '17

11 Jul '17

Avviso di Seminario: Speaker: Rajendra Bhatia Affiliation: Indian Statistical Institute, Delhi Date: July 11 Time: 11:00 -- 12:00 Place: Aula Magna Title: The Bures-Wasserstein distance on positive definite matrices Abstract: We will discuss an interesting metric on the space of positive definite matrices, called the Bures distance in the quantum information literature and the Wasserstein metric in optimal transport. This has connections with Riemannian geometry, statistics, QIT and optimal transport. We will explain these connections. The two-variable and several-variable mean with respect to this metric will be described. Our perspective from matrix analysis leads to simpler proofs and suggests new problems.

2 1

Seminari Benzi e Qi
by Federico Poloni 05 Jun '17

05 Jun '17

Buongiorno, segnalo questo seminario di Michele Benzi in SNS, e ne approfitto per ricordare anche quello di Liqun Qi di lunedì prossimo: > Speaker: Liqun Qi, The Hong Kong Polytechnic University, Department of > Applied Mathematics > > Title: Tensor Analysis: Spectral Theory and Special Tensors > > Room: Sala Riunioni, Dipartimento di Matematica > > Time: Monday, June 12, 11:00 -------- Forwarded Message -------- Subject: [Settimanale] avviso Seminario di Matematica prof. Michele Benzi (8.06.2017) Date: Mon, 5 Jun 2017 10:13:39 +0200 From: Valeria Giuliani <valeria.giuliani(a)sns.it> SEMINARIO DI MATEMATICA Giovedì 8 giugno 2017 ore 11:00 _Scuola Normale Superiore_ Pisa Aula Marie Curie *_ _* * **Michele Benzi*** /(Samuel Candler Dobbs Professor - Emory University)/ */ /* terrà un seminario dal titolo: *“**Decay properties of matrix functions with applications to electronic structure computations**”*** * * *Abstract:* /Functions of matrices, such as the matrix exponential or the square root, arise in a number of scientific and technical applications, ranging from quantum mechanical computations to the structural analysis of complex graphs and networks. The actual numerical evaluation of functions of large and sparse matrices can be quite challenging, and much effort has been devoted to this problem in recent years. An important property that is frequently exhibited by analytic functions of large and sparse matrices is the rapid off-diagonal decay of the matrix entries; careful exploitation of this property can lead to algorithms with optimal computational complexity (this means that the cost scales linearly with the dimension of the problem)./// /In this talk I will present some rigorous estimates on the size of the off-diagonal entries of analytic matrix functions. As a special case, the theory yields a proof of the exponential decay, or "localization", of the density matrix (spectral projector) arising in electronic structure computations for gapped systems, a classical problem in solid state physics and quantum chemistry. In addition to giving a precise mathematical form to W. Kohn's notion of "nearsightedness of electronic matter", these estimates provide a sound basis for the development of linear scaling methods./*//* */ /* Tutti gli interessati sono invitati a partecipare. Classe di Scienze Matematiche e Naturali Valeria Giuliani Scuola Normale Superiore Servizio alla Didattica e Allievi tel. 050 509260 Piazza dei Cavalieri, 7 56126 Pisa E-mail: valeria.giuliani(a)sns.it <mailto:valeria.giuliani@sns.it> E-mail: classi(a)sns.it <mailto:classi@sns.it>

1 0

Fwd: ONE DAY ON NUMERICAL LINEAR ALGEBRA
by Federico Poloni 24 May '17

24 May '17

Ricevo da Dario Bini e inoltro. --- ONE DAY ON NUMERICAL LINEAR ALGEBRA June 14, 2017, Room 2AB40 - Dipartimento di Matematica "Tullio Levi-Civita", Universita' di Padova -Torre Archimede, Via Trieste 63, Padova Nel mese di giugno il Dipartimento di Matematica ospitera' tre reputati ricercatori internazionali nell'ambito dell'Algebra Lineare Numerica, di cui due vincitori delle selezioni Visiting Professor Cariparo 2016 e Visiting Scientist dell'Università di Padova. Si è dunque pensato di raccogliere i loro tre seminari in una giornata dedicata a tali tematiche. I seminari sono: 9:30-10:30 Gérard MEURANT (Paris, France) An optimal Q-OR method for solving nonsymmetric linear systems 11:00-12:00 Claude BREZINSKI (Univ. of Lille, France) Shanks sequence transformations and Anderson acceleration 14:30-15:30 Lothar REICHEL (Kent State University, USA) Generalized Krylov subspace methods for lp-lq minimization with application to image restoration Nella locandina allegata si puo' prendere visione degli abstracts e delle tempistiche di svolgimento dell'evento. La giornata si concludera' con una sessione di commenti e di discussione tra i partecipanti e gli speakers. Potranno pertanto essere sottoposti problemi particolari e richieste di approfondimenti. Tutti gli interessati sono invitati a partecipare. Non è prevista una registrazione, ma per ragioni logistiche, seppure non obbligatoria, sara' utile una potenziale manifestazione di partecipazione inviando una mail a michela(a)math.unipd.it Gli organizzatori M. Redivo-Zaglia, F. Tudisco

1 0

Numerical analysis seminar: L.Q. Qi
by Federico Poloni 11 May '17

11 May '17

Speaker: Liqun Qi, The Hong Kong Polytechnic University, Department of Applied Mathematics Title: Tensor Analysis: Spectral Theory and Special Tensors Room: Sala Riunioni, Dipartimento di Matematica Time: Monday, June 12, 11:00 Abstract: Eigenvalues for tensors were introduced in 2005. Since then, a rich theory on eigenvalues of tensors and special tensors has been developed. There are two main definitions of eigenvalues of tensors. One definition is based upon homogeneous polynomial systems. This definition extends the definition of eigenvalues of matrices. A complex number is an eigenvalue of a ``square'' tensor if and only if it is a root of the characteristic polynomial of that tensor. The sum of all the eigenvalues of that tensor is equal to the sum of the diagonal entries of that tensor, while the product of all the eigenvalues of that tensor is equal to the determinant of that tensor. An eigenvalue of a real tensor with a real eigenvector is called an H-eigenvalue. An even order real symmetric tensor is positive semi-definite if and only if all of its H-eigenvalues are nonnegative. Various tensor eigenvalue inclusion theorems have been developed. Another tensor eigenvalue definition has the merit that the eigenvalues under this definition are invariant under orthonormal transformations. This makes eigenvalues under this definition found their applications in engineering, physics and mechanics. Based upon the tensor eigenvalue theory, theories on special tensors have also been developed. In particular, the Perron-Frobenius theory for irreducible nonnegative tensors and weakly irreducible nonnegative tensors has been developed, and found applications in spectral hypergraph theory and higher order Markov chains. Other special tensors include positive semi-definite tensors, sum-of-squares tensors, completely positive tensors and copositive tensors, etc. Eigenvalues of tensors also found their applications in magnetic resonance imaging, elastic mechanics, liquid crystal study, quantum entanglement and classicality problems, etc. A book “Tensor Analysis: Spectral Theory and Special Tensors'' is published by SIAM in April this year. ==== Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Numerical Analysis Seminar: R. Vandebril
by Federico Poloni 04 Apr '17

04 Apr '17

Giorno: Lunedi' 10 Aprile Ora: 11:00 Luogo: Aula Seminari, Dip.to di Matematica Speaker: Raf Vandebril, KU Leuven Titolo: Eigenvalues and eigenvectors of matrix polynomials === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Seminario Michele Benzi in SNS
by Federico Poloni 21 Mar '17

21 Mar '17

Carissimi, mi segnala Giancarlo Bigi che giovedì prossimo 23 marzo alle ore 14:30 ci sarà un seminario di Michele Benzi alla Scuola Normale: Fast iterative solution methods for the incompressible Navier-Stokes equations Giovedì, 23 Marzo 2017 Ore 14:30 Aula Maria Curie (SNS) Michele Benzi (Emory University) http://www.sns.it/seminari/fast-iterative-solution-methods-incompressible-n… (Abstract su http://www.sns.it/sites/default/files/allegati/15-03-2017/2017_03_23_benzi.…) Arrivederci, -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

workshop a Pisa, "matrix equations and tensor techniques"
by Federico Poloni 01 Feb '17

01 Feb '17

Ciao, il prossimo Febbraio al polo Fibonacci organizziamo uno workshop di due giorni, chiamato matrix equations and tensor techniques. All'edizione precedente a Bologna c'erano diversi partecipanti e speaker anche dall'estero (http://mett15.dm.unibo.it/schedule.html), e ci aspettiamo qualcosa di simile anche quest'anno. Se siete interessati a partecipare, o anche a sottomettere un talk, vi invitiamo a registrarvi (gratuitamente) su http://delone.dm.unipi.it/mett17. La scadenza per sottomettere talk è il 4 dicembre, quella per registrarsi è il 27 gennaio. Un saluto, -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 1

Call for Papers DAPS: Data Mining for Science of Success
by Luca Pappalardo 13 Jul '16

13 Jul '16

*Call for papers* *DAPS2016: The First International Workshop on Data mining for the Analysis of Performance and Success*, in conjunction with the IEEE International Conference on Data Mining (ICDM) Website: http://kdd.isti.cnr.it/daps2016/ Date: December 12, 2016, Barcelona, Spain The increasing availability of Big Data provides an unprecedented opportunity to explore the patterns underlying success. From the strategies followed by successful sportsmen to the emergence of runaway videos on YouTube, from popularity in social media to rising starts in the scientific enterprise, from widespread technologies to groundbreaking innovations, there is wealth of data that can be explored to answer common questions: How can we measure performance? What are the common patterns of success? How did successful individuals and products get to the top? What are the principles driving the dynamics of success? These are the challenging questions at the core of the emerging *Science of Success*, an interdisciplinary field that is attracting scientists from different scientific backgrounds. The purpose of this workshop is to bring together researchers from a variety of areas, all working on the problem of analyzing and understanding the patterns of success, though from different angles. The aim is to discuss: 1) the recently developed machine learning and data mining techniques that can be leveraged to address challenges in analyzing performance and success, and 2) starting from challenges in analyzing performance and success, the practical research directions in the machine learning and data mining community. The best papers of the workshop will be invited to contribute to a special issue on "Performance and Success" edited by the journal Complexity. For more information visit the workshop website: http://kdd.isti.cnr.it/daps2016, and the ICDM conference website: http://icdm2016.eurecat.org/. You can follow news about DAPS2016 on Twitter (@DAPSworkshop) and Facebook (https://www.facebook.com/DAPSworkshop/) *Important Dates:* - *Paper Submission: August 01, 2016, 23:59 PST* - Notification: September 13, 2016 - Camera ready: September 20, 2016 - Workshop: December 12, 2016 *Topics of interest:* We encourage submissions on all theoretical and practical aspects, as well as experimental studies of deployed systems. Topics of interest to the workshop include, but are not limited to, the following subject categories: - Science of Success - Big Data sources for performance and success - mathematical models of success, performance and their interplay - ranking algorithms for performance and success - patterns of success in the scientific enterprise - popularity dynamics of popularity ion social media (hashtags, viral videos, etc.) - definition and analysis of success in social networks - measures of performance and success in sports - data mining for sports analytics - anomaly, doping and fraud detection in sports performance - patterns of success in visual arts, music and writing - adoption, diffusion and success of innovations, products and technologies - novel approaches for detecting innovators and influencers - novel approaches for performance sensing and analysis - design of predictive models for performance and success - opinion mining and sentiment analysis to quantify success in social media - career success and longevity in different professions - analysis of collective success, and well-being, and development - mathematical models of success, performance and their interplay - role of networks in success *Contact information:* For problems or questions, please do not hesitate to contact the workshop organizers by email: - Luca Pappalardo: lpappalardo(a)di.unipi.it - Paolo Cintia: paolo.cintia(a)isti.cnr.it - Roberta Sinatra: sinatrar(a)ceu.edu

1 1

Optimization & Numerical Analysis seminars: Flores-Bazan
by Federico Poloni 23 Jun '16

23 Jun '16

Speaker: Fabián Flores Bazán Affiliation: Departamento de Ingeniería Matemática, Universidad de Concepción, Chile Time and date: Monday, 4th July 2016, 15:00 Place: Sala Seminari Est, Dipartimento di Informatica, Università di Pisa Title: A semi-infinite programming approach for the circular conic StQOP: strong duality, optimality and hidden convexity Abstract: Many formulations of quadratic allocation problems, portfolio optimization problems, the maximum weight clique problem, among others, take the form as the well-known standard quadratic optimization problem, which consist in minimizing a homogeneous quadratic function on the usual simplex in the non negative orthant. We propose to analyze the same problem when the simplex is substituted by a convex and compact base of any pointed, closed, convex (possibly circular) cone. It is associated three main duals (for which a semi-infinite formulation of the primal problem is required) and establish some characterizations of strong duality with respect to each of the three duals in terms of copositivity of the Hessian of the quadratic objective function on suitable cones. Such a problem reveals a hidden convexity. Part of the results were obtained jointly with G. Cárcamo and S. Caro. This research is supported in part by CONICYT-Chile through FONDECYT 115-0973 and BASAL projects, CMM, Universidad de Chile. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Optimization & Numerical Analysis seminars: Romani
by Federico Poloni 20 Jun '16

20 Jun '16

Speaker: Francesco Romani Affiliation: Scuola Normale Superiore, Pisa Time: Monday, 20 June 2016, h. 10:00 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: Algoritmi adattivi per il Clustering convesso e non convesso basato sulla NMF simmetrica Abstract: Vengono presentati algoritmi adattivi basati sulla fattorizzazione simmetrica non negativa di matrici sia per il caso di cluster con inviluppo convesso che di cluster qualsiasi. A parte il calcolo delle distanze tra i punti, gli algoritmi non fanno uso di nozioni geometriche. Lavoro congiunto con Paola Favati (CNR), Grazia Lotti (Università di Parma), Ornella Menchi http://pages.di.unipi.it/fpoloni/files/seminari/romani.doc === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

2 3

Reminder: Seminario Clustering
by Francesco Romani 20 Jun '16

20 Jun '16

Speaker: Francesco Romani Time: Monday, 20 June 2016, h. 10:00 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: Algoritmi adattivi per il Clustering convesso e non convesso basati sulla NMF simmetrica Lavoro congiunto con Paola Favati (CNR), Grazia Lotti (Università di Parma), Ornella Menchi ____________________________ Prof. Francesco Romani Dipartimento di Informatica Via Buonarroti 2, 56127 PISA, ITALY Tel +39-050-2212734 FAX +39-050-2212726 http://www.di.unipi.it/~romani

1 0

Fwd: Reminder: Seminario Clustering
by Francesco Romani 20 Jun '16

20 Jun '16

Speaker: Francesco Romani Time: Monday, 20 June 2016, h. 10:00 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: Algoritmi adattivi per il Clustering convesso e non convesso basati sulla NMF simmetrica Lavoro congiunto con Paola Favati (CNR), Grazia Lotti (Università di Parma), Ornella Menchi

1 0

Reminder: Optimization & Numerical Analysis seminars: Brugiapaglia
by Federico Poloni 20 Jun '16

20 Jun '16

Speaker: Simone Brugiapaglia Affiliation: Politecnico di Milano --- Laboratory for Modeling and Scientific Computing MOX Time: Tuesday, December 15, 11 am Place: Sala Seminari Est, Dipartimento di Informatica, Università di Pisa Title: CORSING: Sparse approximation of PDEs based on Compressed Sensing Abstract: We present a novel method for the numerical approximation of PDEs, motivated by recent developments in sparse representation, and particularly by compressed sensing. We named this approach CORSING (COmpRessed SolvING). Establishing an analogy between the sampling of a signal and the Petrov-Galerkin discretization of a PDE, the CORSING method can recover the best s-term approximation to the solution with respect to N suitable trial functions, with s<<N, by evaluating the bilinear form associated with the PDE against a randomized choice of m<<N test functions. This yields an underdetermined m x N linear system, that is solved by means of sparse optimization techniques. A theoretical analysis of the CORSING procedure is presented, based on the concepts of local a-coherence and restricted inf-sup property, along with numerical experiments that confirm the robustness and reliability of the proposed strategy. === Everyone is welcome! -- --federico poloni

2 8

Fwd: cancellazione seminario
by Dario A. Bini 30 May '16

30 May '16

Mi dispiace dovere annunciare la cancellazione del seminario di Stefano Serra Capizzano previsto oggi alle 11:00. Purtroppo nel viaggio di rientro dagli USA, per un guasto all'aereo per Austin, Stefano ha perso la coincidenza da new York per Roma. Dario

1 0

Avviso di Seminario
by Dario A. Bini 30 May '16

30 May '16

Speaker: Stefano Serra Capizzano, Universita' dell'Insubria Titolo: Spettri, Matrici e Applicazioni nella Approssimazione di Equazioni alle Derivate Parziali Abstract: La presentazione è volta ad illustrare come dedurre informazioni sullo spettro di matrici provenienti dalla approssimazione numerica di equazioni alle derivate parziali: il secondo passo, di interesse squisitamente numerico, consiste nell'impiegare l'informazione ottenuta per il 'design' di metodi efficienti per il calcolo della soluzione dei risultanti sistemi lineari di grandi dimensioni. La chiacchierata, molto informale, includerà una breve discussione su alcuni problemi applicativi. Data: 30 Maggio 2016, ore 11:00 Luogo: Sala riunioni, Dipartimento di Matematica

2 1

Fwd: [Dipartimento.di] Mauriana Pesaresi Seminar Update
by Federico Poloni 11 May '16

11 May '16

Scusate il nuovo messaggio ma c'è un cambio di location per il seminario che vi ho appena inoltrato. -f -------- Forwarded Message -------- Subject: [Dipartimento.di] Mauriana Pesaresi Seminar Update Date: Wed, 11 May 2016 17:34:43 +0200 From: Marco Grandi <marco.grandi(a)di.unipi.it> To: dipartimento(a)di.unipi.it Dear all, the location of tomorrow's Mauriana Pesaresi Seminar has been changed from East Seminar Room to Open Lab. Below updated information on the seminars. Speaker: Giulio Masetti Date: 12 May 2016, 13:00 Place: Open Lab (Room 306) Marco Grandi ______________________________________________ Dipartimento.di mailing list Dipartimento.di(a)listgateway.unipi.it http://listgateway.unipi.it/mailman/listinfo/dipartimento.di

1 0

Fwd: [Dipartimento.di] Mauriana Pesaresi Seminar
by Federico Poloni 11 May '16

11 May '16

Inoltro questo avviso di seminario che sembra di argomento interessante per gli iscritti alla lista. -federico -------- Forwarded Message -------- Subject: [Dipartimento.di] Mauriana Pesaresi Seminar Date: Wed, 11 May 2016 10:53:01 +0200 From: Francesca Pratesi <pratesi(a)di.unipi.it> Organization: University of Pisa To: dottorandi(a)di.unipi.it, dipartimento(a)di.unipi.it Dear all, I would like to remind you that the next Mauriana Pesaresi Seminar will be held tomorrow, 12 May. Below further information on the seminars. Title2: Impact of Equations Ordering on Inexact and Quasi-Inverse Newton Methods. Speaker2: Giulio Masetti Date2: 12 May 2016, 13:00 Place2: East Seminar Room Abstract2: The Newton-Raphson method has a prominent role in applied science and engineering. Our research group at SEDC lab., ISTI/CNR, is involved in dependability-related assessment of Smart Electrical Grids and the Newton-Raphson method is used to solve numerically the system of power-flow equations that characterize the electrical network. In general, given a system of non-linear equations written in the form F(u)=0, where F is a continuously differentiable function that maps a n-dimensional vector u to a n-dimensional vector F(u), one of the solutions can be found via the Newton-Raphson method. At each iteration of this method, a linear system that involves the Jacobian matrix J_{F}(u) of F has to be solved and the equation ordering does not affect it. At increasing the size of the problem, solving with a direct method the linear system is often impractical (w.r.t. time and space complexity, and accuracy), even if the Jacobian is a sparse matrix. Therefore, in these cases iterative methods have been studied, where, the equation ordering becomes a critical aspect to analyse carefully. Studies related to equations ordering are available on techniques employed in medium-size linear systems, such as the Gauss-Seidel method. In this method, at each iteration, permutations of Jacobian rows and columns, that do not depend on the initial equation ordering, are exploited to gain in CPU time. Jacobian-free techniques, developed to solve very large linear problems, offer interesting opportunities for investigations on the impact of initial equations ordering over the number of iterations. One of the most effective algorithms in this category is the Inexact-Newton-Krylov: it gets rid of the Jacobian matrix and approximates the matrix-vector product J_{F}(u)v with the finite difference [F(u+sv)-F(u)]/s. Another variant of Newton- Raphson is the Quasi-Inverse-Newton: it approximates the inverse of the Jacobian J_{F}^{-1}(u) with a diagonal matrix and solves directly the linear system. A seminal article, written by L. Dutto and published on the International Journal for Numerical Methods in Engineering in 1993, numerically presents the influence of initial ordering over the preconditioner of Inexact- Newton-Krylov GMRES, and then on the number of iterations. However, it does not address the impact of ordering on the core of the method, that is the adopted Jacobian-free techniques. To go further such existing results, our study focuses on Jacobian- free methods and explores the influence of the initial ordering over them. In this talk, we discuss our approach to prove that the initial equations ordering affects Inexact-Newton-Krylov and Quasi- Inverse-Newton, also presenting a preliminary analysis to quantify the resulting impact. A trivial bivariate function will be used as case study and numerical experiments will be presented. Interestingly, the equations ordering does not only affect the efficiency in reaching a solution, but also the solution itself. As next step, we are working on exploring more realistic scenarios in the power grid domain. Traditionally, the characteristics of the domain at hand has influenced the application of mathematical structures. For example, when using Gauss-Seidel in the power-flow context, there exists a particular initial ordering such that the Jacobian is a four-block matrix and each block corresponds to well defined electrical quantities. Using Jacobian-free methods the application-oriented initial equations ordering could not be the best choice with respect to iterations number and reachable solutions. We believe this is an important feature, that the user of these methods need to be aware of. -- Francesca Pratesi Ph.D. Student - University of Pisa Room 296a - Phone 050 2212731 Department of Computer Science, Largo B. Pontecorvo 3, 56127 Pisa ______________________________________________ Dipartimento.di mailing list Dipartimento.di(a)listgateway.unipi.it http://listgateway.unipi.it/mailman/listinfo/dipartimento.di

1 0

Funerali
by Francesco Romani 09 May '16

09 May '16

Il funerali della madre di Grazia Lotti sono domani alle 11 alla cappella del cimitero della Misericordia. Francesco. ____________________________ Prof. Francesco Romani Dipartimento di Informatica Via Buonarroti 2, 56127 PISA, ITALY Tel +39-050-2212734 FAX +39-050-2212726 http://www.di.unipi.it/~romani

1 0

Optimization & Numerical Analysis seminars: Massei
by Federico Poloni 27 Apr '16

27 Apr '16

(This message also serves as a reminder that Claudio Sterle's seminar on "Flow intercepting facility location routing problem: formulation and branch-and-cut solving approach" is scheduled in two days: Friday, 29 April 2016, h. 14:30. ) Speaker: Stefano Massei Affiliation: Scuola Normale Superiore, Pisa Time: Thursday, 5 May 2016, h. 10:00 Place: Sala Seminari Est, Dipartimento di Informatica, Università di Pisa Title: Numerical quasiseparable preservation in matrix functions and in the cyclic reduction Abstract: The quasiseparable structure can be efficiently exploited - in order to speed up computations and storage - by using hierachical representations. This theme has raised much interest in the last two decades, in particular in the context of elliptic partial differential equations, matrix functions and Sylvester equations. Our aim is to provide a framework for the analysis of the off-diagonal singular values, focusing on the structures that can ensure an exponential decay of them. We see how this structures appears in the computation of matrix functions and in the execution of the cyclic reduction paving the way for a theoretical understanding of the preservation of the quasiseparable rank. We provide family of bounds that highlight which parameters of the model can affect the rate of decay. Finally, we show how to practically exploit the low quasiseparable rank using hierarchical matrices. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Optimization & Numerical Analysis seminars: Sterle
by Federico Poloni 14 Apr '16

14 Apr '16

Speaker: Claudio Sterle Affiliation: University “Federico II” of Naples Time: Friday, 29 April 2016, h. 14:30 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: Flow intercepting facility location routing problem: formulation and branch-and-cut solving approach Abstract: In the last ten years the research activity on location-routing problem (LRP) has significantly increased and many effective exact and approximated approaches have been presented for its solution. Following from this, great attention has been also devoted to several variants of the LRP, most of them coming from real applications. In particular, with reference to the city logistic (CL) context, several variants of the LRP have been proposed to design single-tier and two-tier freight distribution systems. In this case, LRP location decisions concern the number and position of one or more types of logistic facilities, whereas, routing decisions, concern the definition of dedicated or multi-stop routes of homogeneous or heterogeneous vehicle fleets. However most of the present literature deals with single-commodity LRP, whereas the multi-commodity case is still scarcely investigated. This issue can assume an important role in LRP real applications and in particular in the CL context. Indeed the freight to be distributed in an urban area are highly heterogeneous and consequently the problem is inherently multi-commodity. In order to contribute to fill this gap, an original formulation for the multi-commodity LRP is presented. The problem is referred to as the flow intercepting facility location-routing problem (FIFLOR), since its formulation is obtained merging the flow intercepting facility location and the vehicle routing models. A branch-and-cut algorithm for its solution is described and results obtained on test instances reproducing different urban network topologies are discussed. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Reminder: Optimization and numerical analysis seminars
by Federico Poloni 08 Apr '16

08 Apr '16

Speaker: Yaroslav D. Sergeyev Affiliation: Università della Calabria and N.I. Lobachevsky University of Nizhni Novgorod, Russia. Time: Monday, 11 April 2016, 15:00 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: Lipschitz global optimization Abstract: http://pages.di.unipi.it/fpoloni/files/seminari/Sergeyev.pdf and Time: Tuesday, 12 April 2016, 12:00 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: The Infinity Computer and numerical computations with infinities and in infinitesimals Abstract: http://pages.di.unipi.it/fpoloni/files/seminari/Sergeyev_Infinity.pdf === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Cena di giovedì 7 Aprile
by massei＠student.dm.unipi.it 05 Apr '16

05 Apr '16

Cari tutti, in occasione della partenza di Raf Vandebril e come scusa per stare un po' insieme, si pensava di fare una cena giovedì sera alla trattoria di Montemagno [1]. Chi ha piacere di partecipare è ben accetto. Dovreste però comunicarmi chi viene e se ha un mezzo a disposizione, così prenoto e ci organizziamo per raggiungere il posto. A presto e buona giornata, Stefano [1] https://www.tripadvisor.it/Restaurant_Review-g580236-d1867657-Reviews-Tratt…

1 0

Optimization & Numerical Analysis seminars: Sergeyev
by Federico Poloni 30 Mar '16

30 Mar '16

Speaker: Raf Vandebril Affiliation: KU Leuven Time: Thursday, April 7, h. 10:00 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: A framework for structured linearizations of matrix polynomials in various bases Abstract: We present a framework for the construction of linearizations for scalar and matrix polynomials based on dual bases which, in the case of orthogonal polynomials, can be described by the associated recurrence relations. The framework provides an extension of the classical linearization theory for polynomials expressed in non-monomial bases and allows to represent polynomials expressed in product families, that is, as a linear combination of elements of the form $\phi_i(x) \psi_j(x)$, where $\{ \phi_i(x) \}$ and $\{ \psi_j(x) \}$ can either be polynomial bases or polynomial families which satisfy some mild assumptions. We show that this general construction can be used for many different purposes. Among them, we show how to linearize sums of polynomials and rational functions expressed in different bases. As an example, this allows to look for intersections of functions interpolated on different nodes without converting them to the same basis. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Optimization & Numerical Analysis seminars: Sergeyev (II)
by Federico Poloni 24 Mar '16

24 Mar '16

Speaker: Yaroslav D. Sergeyev Affiliation: Università della Calabria and N.I. Lobachevsky University of Nizhni Novgorod, Russia. Time: Tuesday, 12 April 2016, 12:00 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: The Infinity Computer and numerical computations with infinities and in infinitesimals Abstract: The lecture presents a recent methodology allowing one to execute numerical computations with finite, infinite, and infinitesimal numbers on a new type of a computer – the Infinity Computer – patented in EU, USA, and Russia (see [23]). The new approach is based on the principle ‘The whole is greater than the part’ (Euclid’s Common Notion 5) that is applied to all numbers (finite, infinite, and infinitesimal) and to all sets and processes (finite and infinite). It is shown that it becomes possible to write down finite, infinite, and infinitesimal numbers by a finite number of symbols as particular cases of a unique framework different from that of the non-standard analysis [...] http://pages.di.unipi.it/fpoloni/files/seminari/Sergeyev_Infinity.pdf === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Optimization & Numerical Analysis seminars: Sergeyev
by Federico Poloni 17 Mar '16

17 Mar '16

Speaker: Yaroslav D. Sergeyev Affiliation: Università della Calabria and N.I. Lobachevsky University of Nizhni Novgorod, Russia. Time: Monday, 11 April 2016, 15:00 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: Lipschitz global optimization Abstract: Global optimization is a thriving branch of applied mathematics and an extensive literature is dedicated to it (see e.g., [1–24]). In this lecture, the global optimization problem of a multidimensional function satisfying the Lipschitz condition over a hyperinterval with an unknown Lipschitz constant is considered. It is supposed that the objective function can be “black box”, multiextremal, and non-differentiable. It is also assumed that evaluation of the objective function at a point is a time-consuming operation. Many algorithms for solving this problem have been discussed in literature. They can be distinguished, for example, by the way of obtaining information about the Lipschitz constant and by the strategy of exploration of the search domain. Different exploration techniques based on various adaptive partition strategies are analyzed. [...] http://pages.di.unipi.it/fpoloni/files/seminari/Sergeyev.pdf === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Seminari su Mathematica
by Francesco Romani 02 Mar '16

02 Mar '16

Nell’ambito del corso di Esperienze di programmazione saranno tenuti 4 seminari introduttivi di circa 2 ore sul linguaggio Mathematica. Mercoledì 9/3 11-13 Aula C1 Introduzione a Mathematica Mercoledì 16/3 11-13 Aula C1 Grafica con Mathematica Mercoledì 23/3 11-13 Aula C1 Programmazione in Mathematica Mercoledì 27/4 11-13 Aula C1 Calcolo Numerico con Mathematica La partecipazione è aperta a tutti e vi prego di diffondere l’informazione ad eventuali interessati. Grazie, saluti, FR ____________________________ Prof. Francesco Romani Dipartimento di Informatica Via Buonarroti 2, 56127 PISA, ITALY Tel +39-050-2212734 FAX +39-050-2212726 http://www.di.unipi.it/~romani

1 0

Mathematica
by Francesco Romani 29 Feb '16

29 Feb '16

Anche quest’anno, all’interno del corso di Esperienze di Programmazione, faccio 4-5 seminari su Mathematica. L’orario è mercoledi 11-13 / Venerdi 16-18 Avendo una certra libertà per collocarli volevo sapere se qualcuno ha intenzione di seguirli e ha preferenze sugli orari. Grazie saluti Francesco ____________________________ Prof. Francesco Romani Dipartimento di Informatica Via Buonarroti 2, 56127 PISA, ITALY Tel +39-050-2212734 FAX +39-050-2212726 http://www.di.unipi.it/~romani

1 0

Optimization & Numerical Analysis seminars: Masciotta
by Federico Poloni 29 Feb '16

29 Feb '16

Speaker: Andrea Masciotta Affiliation: Dipartimento di Matematica, Università di Pisa Time: Monday, 7 March 2016, 14:30 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: Minimization algorithms for compressive sensing Abstract: Compressive sensing is a new approach to digital signals acquisition and processing. Its main aim is the reconstruction of sparse signals or images (i.e. with few significant coefficients) from what was previously believed to be incomplete information. Compressive sensing has many potential applications, not only limited to the analysis of signals or images. In this talk we will first describe the main theoretical properties of compressive sensing, then three solution algorithms based on \ell_1 minimization will be presented, along with numerical experiments to highlight the main characteristics of these three methods in the framework of compressive sensing. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Optimization & Numerical Analysis seminars: Miglierina
by Federico Poloni 17 Feb '16

17 Feb '16

Speaker: Enrico Miglierina Affiliation: Università Cattolica del Sacro Cuore, Milano Time: Thursday, 25 February 2016, 15:00 Place: Sala Seminari Est, Dipartimento di Informatica, Università di Pisa Title: Stability by set convergences: from Vector to Set optimization Abstract: (pdf) http://pages.di.unipi.it/fpoloni/files/seminari/Miglierina.pdf The notions of set convergences reveal to be very useful to study stability properties in vector optimization. Indeed, some results that show how the convergence of feasible region implies the convergence of efficient frontiers are known. The oldest ones are based on suitable compactness assumptions that concern the sequence of sets as a whole(see \cite{Attouch-Rihai,Bednarczuk2007,Luc-Lucchetti-Malivert,LoridanMorganRaucci,Penot-Sterna}). Moreover, more recently, some results about convergence of efficient frontiers have been obtained by replacing compactness by convexity (\cite{Lucchetti-Miglierina2004Opt,MM2005}). After a presentation of these results, the last aim of the talk is to extend this class of results to the broader framework of set optimization, that recently received an increasing attention in the literature (\cite{Gutierrez-Miglierina-Molho-Novo}). === Everyone is welcome! -- --federico poloni _______________________________________________ Numpi.di mailing list Numpi.di(a)listgateway.unipi.it http://listgateway.unipi.it/mailman/listinfo/numpi.di

1 0

[Dipartimento.di] Reminder: Optimization & Numerical Analysis seminars: Strabic
by Federico Poloni 18 Dec '15

18 Dec '15

Speaker: Nataša Strabić Affiliation: The University of Manchester Time: Wednesday, December, 9; h. 15:00 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: Recent Progress on the Nearest Correlation Matrix Problem Abstract: In a wide range of applications it is required to replace an empirically obtained unit diagonal indefinite symmetric matrix with a valid correlation matrix (unit diagonal positive semidefinite matrix). A popular replacement is the nearest correlation matrix in the Frobenius norm. The first method for computing the nearest correlation matrix with guaranteed convergence was the alternating projections method proposed by Higham in 2002. The rate of convergence of this method is at best linear, and it can require a large number of iterations to converge to within a given tolerance. Although a faster globally convergent Newton algorithm was subsequently developed by Qi and Sun in 2006, the alternating projections method remains very widely used. We show that Anderson acceleration, a technique for accelerating the convergence of fixed-point iterations, can be applied to the alternating projections method and that in practice it brings a significant reduction in both the number of iterations and the computation time. We also show that Anderson acceleration remains effective, and indeed can provide even greater improvements, when it is applied to the variants of the nearest correlation matrix problem in which specified elements are fixed or a lower bound is imposed on the smallest eigenvalue. This is particularly significant for the nearest correlation matrix problem with fixed elements because no Newton method with guaranteed convergence is available for it. Both methods for computing the nearest correlation matrix are based on repeated eigenvalue decompositions and so they can be infeasible in time-critical situations. We have recently proposed an alternative method to restore definiteness to an indefinite matrix called shrinking. The method is based on computing the optimal parameter in a convex linear combination of the indefinite starting matrix and a chosen positive definite target matrix. We show how this problem can be solved by the bisection method and posed as a generalized eigenvalue problem, and we demonstrate how exploiting positive definiteness in these two methods leads to impressive computational savings. The work on these two topics is joint with Nicholas J. Higham, and, for shrinking, with Vedran Šego. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel. +39 050 2213143 ______________________________________________ Dipartimento.di mailing list Dipartimento.di(a)listgateway.unipi.it http://listgateway.unipi.it/mailman/listinfo/dipartimento.di

2 1

[Dipartimento.di] Optimization & Numerical Analysis seminars: Strabic
by Federico Poloni 03 Dec '15

03 Dec '15

Speaker: Nataša Strabić Affiliation: The University of Manchester Time: Wednesday, December, 9; h. 15:00 Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: Recent Progress on the Nearest Correlation Matrix Problem Abstract: In a wide range of applications it is required to replace an empirically obtained unit diagonal indefinite symmetric matrix with a valid correlation matrix (unit diagonal positive semidefinite matrix). A popular replacement is the nearest correlation matrix in the Frobenius norm. The first method for computing the nearest correlation matrix with guaranteed convergence was the alternating projections method proposed by Higham in 2002. The rate of convergence of this method is at best linear, and it can require a large number of iterations to converge to within a given tolerance. Although a faster globally convergent Newton algorithm was subsequently developed by Qi and Sun in 2006, the alternating projections method remains very widely used. We show that Anderson acceleration, a technique for accelerating the convergence of fixed-point iterations, can be applied to the alternating projections method and that in practice it brings a significant reduction in both the number of iterations and the computation time. We also show that Anderson acceleration remains effective, and indeed can provide even greater improvements, when it is applied to the variants of the nearest correlation matrix problem in which specified elements are fixed or a lower bound is imposed on the smallest eigenvalue. This is particularly significant for the nearest correlation matrix problem with fixed elements because no Newton method with guaranteed convergence is available for it. Both methods for computing the nearest correlation matrix are based on repeated eigenvalue decompositions and so they can be infeasible in time-critical situations. We have recently proposed an alternative method to restore definiteness to an indefinite matrix called shrinking. The method is based on computing the optimal parameter in a convex linear combination of the indefinite starting matrix and a chosen positive definite target matrix. We show how this problem can be solved by the bisection method and posed as a generalized eigenvalue problem, and we demonstrate how exploiting positive definiteness in these two methods leads to impressive computational savings. The work on these two topics is joint with Nicholas J. Higham, and, for shrinking, with Vedran Šego. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel. +39 050 2213143 ______________________________________________ Dipartimento.di mailing list Dipartimento.di(a)listgateway.unipi.it http://listgateway.unipi.it/mailman/listinfo/dipartimento.di

1 0

Optimization & Numerical Analysis seminars: Fenu
by Federico Poloni 19 Nov '15

19 Nov '15

Speaker: Caterina Fenu Affiliation: Università di Pisa Time: Thursday, November, 26; 9 am Place: Sala Seminari Ovest, Dipartimento di Informatica, Università di Pisa Title: Fast Computation of Centrality Indices Abstract: One of the main issues in complex networks theory is to find the “most important” nodes. To this aim, one can use matrix functions applied to its adjacency matrix. After an introduction on the use of Gauss-type quadrature rules, we will discuss a new computational method to rank the nodes of both directed and undirected (unweighted) networks according to the values of these functions. The algorithm uses a low-rank approximation of the adjacency matrix, then Gauss quadrature is used to refine the computation. Extended pdf abstract on http://www.di.unipi.it/~fpoloni/files/seminari/Fenu.pdf === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel. +39 050 2213143

1 0

Optimization & Numerical Analysis seminars: Brugiapaglia
by Federico Poloni 23 Oct '15

23 Oct '15

Speaker: Simone Brugiapaglia Affiliation: Politecnico di Milano --- Laboratory for Modeling and Scientific Computing MOX Time: Tuesday, December 15, 11 am Place: Sala Seminari Est, Dipartimento di Informatica, Università di Pisa Title: CORSING: Sparse approximation of PDEs based on Compressed Sensing Abstract: We present a novel method for the numerical approximation of PDEs, motivated by recent developments in sparse representation, and particularly by compressed sensing. We named this approach CORSING (COmpRessed SolvING). Establishing an analogy between the sampling of a signal and the Petrov-Galerkin discretization of a PDE, the CORSING method can recover the best s-term approximation to the solution with respect to N suitable trial functions, with s<<N, by evaluating the bilinear form associated with the PDE against a randomized choice of m<<N test functions. This yields an underdetermined m x N linear system, that is solved by means of sparse optimization techniques. A theoretical analysis of the CORSING procedure is presented, based on the concepts of local a-coherence and restricted inf-sup property, along with numerical experiments that confirm the robustness and reliability of the proposed strategy. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Seminario di Dierk Schleicher
by Dario A. Bini 16 Oct '15

16 Oct '15

Speaker: Dierk Schleicher Title: Newton’s method as an unexpectedly efficient root finder Time: Friday, October 23; 10:00 Place: Aula Magna, Dipartimento di Matematica Abstract: Newton’s method is well known as a root finder locally near the roots. It is often “not recommended” as a global root finder because of its “chaotic” properties. We give a very efficient theoretical upper bound on its speed of convergence: all roots of a degree d polynomial can be found with accuracy eps in O(d^2 log^4 d + d log|\log eps|) Newton iterations in the expected case — very close to the theoretical upper bound of O(d^2) for this method. In practice, all roots of polynomials of degree more than a million are found routinely and in (4log 2) d^2 iterations — in practice, this means less than a day for degree a million on standard laptops. A modified method, for which we do not yet have a complete theory, brings this down to 3 d log^2 d iterations, which in practice for certain polynomials finds all roots of degree a million in a matter of minutes. === Everyone is welcome! Dario A. Bini

1 0

Optimization & Numerical Analysis Seminars: Nesti
by Federico Poloni 14 Sep '15

14 Sep '15

Speaker: Tommaso Nesti (University of Pisa) Title: Numerical methods for computing the steady-state distribution of a G-network. Time: Wednesday, September 23; 16:00 Place: Sala seminari est, informatica Abstract: G-networks are a class of queueing networks introduced by E. Gelembe in 1989, which are characterized by the presence of positive and negative customers. Unlike classical queueing networks, the equations yielding the steady-state distribution of a G-network are nonlinear, resulting in the need of developing efficient numerical methods for their solution. In this talk we present two numerical methods for the computation of the steady-state distribution, namely a fixed point and a Newton-Raphson iteration. Using tools from nonnegative matrix theory, we prove that both methods are locally convergent to the fixed point, with linear and quadratic rate of convergence, respectively. We compare these methods with an existing algorithm, concluding that the Newton-Raphson iteration is a preferable choice for G-networks with a moderate number of queues. We conclude by suggesting a possible application for future research. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel. +39 050 2213143

1 0

Evento AIRO (e conferenza relativa)
by Federico Poloni 31 Aug '15

31 Aug '15

Segnalo anche su questa lista l'evento sociale della conferenza AIRO: http://www.airo.org/conferences/airo2015/e-sport Si tratta di un torneo informale di un videogioco open source che si svolgerà al museo degli strumenti per il calcolo la sera dell'8 settembre. "The competition is open to all the conference participants, the accompanying persons, the organizers, any of their relative, friend and acquaintance (recursively), and whomever can claim any relationship with the conference, however tenuous and imaginative it may be. Fabulous final prices, and gadgets for all participants!" L'evento è collegato alla conferenza dell'AIRO (associazione italiana ricerca operativa) che si svolgerà la prossima settimana al dipartimento di informatica http://www.airo.org/conferences/airo2015/program. Per chi è interessato all'algebra lineare numerica segnalo in particolare il minisimposio "Numerical Methods for Complex Networks", in cui parleranno Francesca Arrigo, Dario Fasino e Caterina Fenu. -federico -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Optimization & Numerical Analysis Seminars: Flores-Bazan and Locatelli
by Federico Poloni 29 Jun '15

29 Jun '15

Two seminars in optimization, one after the other, are taking place next Tuesday. Speaker1: Fabian Flores-Bazan (University of Concepcion) Title1: Strong duality in nonconvex optimization with a single equality and geometric constraints: applications to QP Time1: Tuesday, June 30; 15:00 Place1: Sala seminari est, informatica Speaker2: Marco Locatelli (Università di Parma) Title2: Some results about convex envelopes Time2: Tuesday, June 30; 15:45 Place2: Sala seminari est, informatica Abstract1: The validity of strong duality for a non convex optimization problem under a single equality and geometric constraints is characterized in a topological and geometric manner. In particular, a hidden convexity of the conic hull of joint-range of the pair of functions associated to the original problem, is obtained. Applications to derive KKT conditions without standard constraints qualification are also discussed. Several examples showing our results provide much more information than those appearing elsewhere, are given. Furthermore, first and second order optimality condition, for a non convex quadratic optimization problem under two quadratic equality constraints, are also presented. Finally, the standard quadratic problem involving a non necessarily polyhedral cone is analyzed in detail. The talk is based on joints works with G. Carcamo and F. Opazo. Abstract2: In this talk we discuss convex envelopes for some bivariate functions over general polytopes and for some $n$-dimensional quadratic functions over the unit simplex. We first give two formulations of convex envelope, a primal and a dual one. Next, we discuss how to derive the convex envelope for some bivariate functions over polytopes, provided that some given assumptions are fulfilled. It will turn out that the formula of the convex envelope can be derived through the solution of a parametric convex program or, equivalently, through the solution of the corresponding KKT system. For what concerns the $n$-dimensional quadratic functions over the unit simplex, we will first associate a graph to the quadratic function, and next we will show that the derivation of the convex envelope is related to the identification of the spanning forests over this graph. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel. +39 050 2213143

2 1

Fwd: Call per data analyst
by Federico Poloni 18 Jun '15

18 Jun '15

Ricevo e inoltro questa offerta di lavoro. -federico -------- Forwarded Message -------- in Zanichelli cercano un data analyst junior a tempo determinato, e una parte della selezione (quella "tecnica") la faccio io. In allegato c'è la call: potreste girarla, per favore, a chiunque possa essere interessato? Gli interessati possono inviare cv e lettera di presentazione al nevio.dubbini(a)gmail.com. (NB il mio indirizzo email non c'è nella call allegata, quindi andrebbe comunicato nel testo dell'email). Grazie, ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Nevio Dubbini mail: nevio.dubbini(a)gmail.com web: sites.google.com/site/neviodubbini ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

1 0

Fwd: [Dipartimento.di] avviso ciclo di seminari - 8, 10, 11 giugno
by Federico Poloni 04 Jun '15

04 Jun '15

In the framework of the research project "Mathematical models and computational methods for complex networks" by the University of Pisa, Paola Boito (Faculté des Sciences et Techniques Université de Limoges) will give three seminars on "Complex networks, functions of matrices and quadrature rules". The abstract is provided below. The seminars will take place in June at the Department of Computer Science of the University of Pisa: -- June 8th, 15:00-16:00, sala seminari est (room 351); -- June 10th, 10:00-11:00, sala seminari est (room 351); -- June 11th, 15:00-16:00, sala seminari est (room 351). Abstract. One cornerstone of our modern society is the use of different kinds of networks. Our cities are connected by a network of streets and railways, telecommunication networks including their wireless, mobile components, the internet, and the World Wide Web build the most important infrastructure for communication and information worldwide. Designing and managing such networks pose challenging algorithmic and numerical problems. These three talks are concerned with the development, analysis and application of numerical linear algebra tools for the quantitative study of important properties of large-scale complex networks, such as centrality, betweenness and communicability measures. Specifically, we design efficient algorithms for the fast approximation of the entries of matrix functions such as the exponential and the resolvent of adjacency matrices and graph Laplacians, as well as their traces. While these matrices are quite sparse, in the case of complex networks they behave very differently from the matrices associated with regular lattices (grids), such as those arising from discretizations of partial differential equations. Hence, there is a need to develop new methods, both computational and analytical, to deal with these very large-scale problems. Potential applications include the analysis of social networks, the study of biological and neurological networks, applications in physics and operations research, and so forth.

1 1

Optimization & Numerical Analysis Seminar: N. Vannieuwenhoven
by Federico Poloni 03 Jun '15

03 Jun '15

Speaker: Nick Vannieuwenhoven (KU Leuven) Title: Towards a perturbation theory for the tensor rank decomposition Date: Thursday, 4 June 2015. Time: 11:00 Place: Sala Seminari Ovest, dipartimento di informatica Abstract: The tensor rank decomposition is a decomposition of a tensor into a linear combination of rank-1 tensors. One of the key advantages of higher-order tensors is that this decomposition is generally unique, which allows for an interpretation of the individual rank-1 terms. Because of this property, the tensor rank decomposition has found application in several domains. For instance, they can be used as a clustering algorithm in an unsupervised learning setting assuming a certain statistical model wherein each of the rank-1 tensors in the decomposition will correspond with a cluster. In the applications where tensor rank decompositions arise, the tensor is usually known only up to some small perturbation error. This raises the following interesting problem: Suppose that the tensor is perturbed by a small quantity, by how much can the individual rank-1 terms change relative to the magnitude of the perturbation? Can it be unbounded? In this presentation, I will explore the definition of a possible condition number that answers these questions up to first order. == Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 1

Reminder: Optimization & Numerical Analysis Seminars: F. Romani
by Federico Poloni 13 May '15

13 May '15

Date: Thursday May 14, 2015 Time: 11:00 Place: Sala Seminari Ovest, dipartimento di informatica Speaker: Francesco Romani, dipartimento di informatica, U Pisa. Title: *Krylov subspace methods for the solution of linear systems* Abstract: Krylov subspace methods are iterative methods for the solution of linear systems which operate by multiplying the system matrix (sometimes also its transpose) by a vector. They are recommended when the system matrix is large and sparse or with a strong structure. The most common Krylov subspace methods are presented focusing the attention mainly on the practical and empirical aspects. == The seminar will be in Italian (but with slides in English). Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Optimization & Numerical Analysis Seminars: F. Romani
by Federico Poloni 07 May '15

07 May '15

Date: Thursday May 14, 2015 Time: 11:00 Place: Sala Seminari Ovest, dipartimento di informatica Speaker: Francesco Romani, dipartimento di informatica, U Pisa. Title: *Krylov subspace methods for the solution of linear systems* Abstract: Krylov subspace methods are iterative methods for the solution of linear systems which operate by multiplying the system matrix (sometimes also its transpose) by a vector. They are recommended when the system matrix is large and sparse or with a strong structure. The most common Krylov subspace methods are presented focusing the attention mainly on the practical and empirical aspects. == The seminar will be in Italian (but with slides in English). Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Reminder: Optimization & Numerical Analysis Seminar: M.G. Scutellà
by Federico Poloni 29 Apr '15

29 Apr '15

Our next seminar is tomorrow morning. Date: Thursday April 30, 2015 Time: 10:00 Place: Sala Seminari Ovest, dipartimento di informatica Speaker: Maria Grazia Scutellà, dipartimento di informatica, U Pisa. Title: *Optimizing Green Wireless LANs* Abstract: We consider problems arising in the design of Green (or energy-saving) Wireless Local Area Networks (GWLANs). In this context, decisions on powering-on a set of access points, and decisions on the assignment of the user terminals to the opened access points, have to be taken simultaneously. In particular, the level of power assigned to each access point may affect, in a nonlinear way, the capacity of the connections between the access points and the user terminals that are assigned to them. The aim is to minimize the overall power consumption of the access points, which is given by a linear dependency between the power consumed and the total demand assigned to the access points, via assignment costs. In the first part of the talk we present a Mixed Integer formulation with NonLinear constraints (MINLP) and an optimization approach, belonging to the framework of Branch and Benders Cut methods, which is based on the proposed MINLP formulation. In a non-standard fashion, the master problem of the approach includes the variables of the Benders subproblem, but relaxes their integrality. The approach has been tested on a large set of instances, and compared to a more traditional Benders decomposition algorithm on a subset of instances without the assignment costs, where the two approaches can be compared. The main achievements of the computational experimentation are shortly discussed. Then, in the second part of the talk, we address the optimization of GWLANs in a more realistic scenario, which takes into account also some uncertainty factors. Specifically, we present a robust optimization approach that incorporates both link capacity fluctuations and user mobility under Bertsimas and Sim's robust optimization paradigm. Again, preliminary computational results are shortly discussed. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel. +39 050 2213143

1 0

Optimization & Numerical Analysis Seminar: M.G. Scutellà
by Federico Poloni 23 Apr '15

23 Apr '15

Date: Thursday April 30, 2015 Time: 10:00 Place: Sala Seminari Ovest, dipartimento di informatica Speaker: Maria Grazia Scutellà, dipartimento di informatica, U Pisa. Title: *Optimizing Green Wireless LANs* Abstract: We consider problems arising in the design of Green (or energy-saving) Wireless Local Area Networks (GWLANs). In this context, decisions on powering-on a set of access points, and decisions on the assignment of the user terminals to the opened access points, have to be taken simultaneously. In particular, the level of power assigned to each access point may affect, in a nonlinear way, the capacity of the connections between the access points and the user terminals that are assigned to them. The aim is to minimize the overall power consumption of the access points, which is given by a linear dependency between the power consumed and the total demand assigned to the access points, via assignment costs. In the first part of the talk we present a Mixed Integer formulation with NonLinear constraints (MINLP) and an optimization approach, belonging to the framework of Branch and Benders Cut methods, which is based on the proposed MINLP formulation. In a non-standard fashion, the master problem of the approach includes the variables of the Benders subproblem, but relaxes their integrality. The approach has been tested on a large set of instances, and compared to a more traditional Benders decomposition algorithm on a subset of instances without the assignment costs, where the two approaches can be compared. The main achievements of the computational experimentation are shortly discussed. Then, in the second part of the talk, we address the optimization of GWLANs in a more realistic scenario, which takes into account also some uncertainty factors. Specifically, we present a robust optimization approach that incorporates both link capacity fluctuations and user mobility under Bertsimas and Sim's robust optimization paradigm. Again, preliminary computational results are shortly discussed. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel. +39 050 2213143

1 0

Reminder: Optimization & Numerical Analysis Seminar: F. Poloni
by Federico Poloni 16 Apr '15

16 Apr '15

Date: Thursday April 16, 2015 Time: 11:00 Place: Sala Riunioni Est, dipartimento di informatica Speaker: Federico Poloni, dipartimento di informatica, U Pisa. Title: *Methods for accurate invariant subspace computations.* Abstract: We consider two modelling problems, one arising in control theory (linear-quadratic regulator) and one arising in the modelling of queues (or buffers) under a continuous-time Markov model. The solution of both can be reduced to finding the invariant subspace of a suitable structured matrix, or equivalently solving a Riccati-like matrix equation. I will present linear algebra tools that can be used to solve them accurately; in particular: * a QR-like matrix factorization in which a tall skinny n x m matrix is factored as U*R, where R is m x m square invertible and U has a m x m identity submatrix and all its other entries bounded in modulus by 1. This factorization can be formulated as an optimization problem (maximum volume submatrix) or as finding a suitable set of basic variables in a tableau. * (if time allows) a numerical method for the Markov problem based on a technique (GTH method) to solve linear systems with a (possibly ill-conditioned) M-matrix without subtractions, obtaining high accuracy in both large and small entries of the solution. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel. +39 050 2213143

1 0

Optimization & Numerical Analysis Seminar: F. Poloni
by Federico Poloni 10 Apr '15

10 Apr '15

Date: Thursday April 16, 2015 Time: 11:00 Place: Sala Riunioni Est, dipartimento di informatica Speaker: Federico Poloni, dipartimento di informatica, U Pisa. Title: *Methods for accurate invariant subspace computations.* Abstract: We consider two modelling problems, one arising in control theory (linear-quadratic regulator) and one arising in the modelling of queues (or buffers) under a continuous-time Markov model. The solution of both can be reduced to finding the invariant subspace of a suitable structured matrix, or equivalently solving a Riccati-like matrix equation. I will present linear algebra tools that can be used to solve them accurately; in particular: * a QR-like matrix factorization in which a tall skinny n x m matrix is factored as U*R, where R is m x m square invertible and U has a m x m identity submatrix and all its other entries bounded in modulus by 1. This factorization can be formulated as an optimization problem (maximum volume submatrix) or as finding a suitable set of basic variables in a tableau. * (if time allows) a numerical method for the Markov problem based on a technique (GTH method) to solve linear systems with a (possibly ill-conditioned) M-matrix without subtractions, obtaining high accuracy in both large and small entries of the solution. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel. +39 050 2213143

1 0

Reminder: Optimization & Numerical Analysis seminars: R. Vandebril
by Federico Poloni 08 Apr '15

08 Apr '15

Date: Giovedi' 9 Aprile Time: 11:00 Place: Aula Seminari, dipartimento di matematica Speaker: Raf Vandebril, Department of Computer Science, KU Leuven. Title: *The Companion QR* Abstract: In this lecture we will propose a new fast and stable manner of computing roots of polynomials. Roots of polynomials are typically computed by putting the coefficients of the polynomial in the companion matrix and then computing the eigenvalues of this matrix. Exploiting the available low-rank structure leads to an algorithm of quadratic instead of cubic complexity. Several low-cost algorithms have already been proposed. Either they fully exploit the low-rank properties of the involved matrices by representing them for instance via quasiseparable factors, or they write the companion matrix as the sum of a unitary plus low rank matrix and exploit this structure. In this lecture we will use the second manner. However, only a single QR step requires the use of the low rank part. After that we can put the low rank term aside and continue only with the unitary matrix, translating the problem thereby to a unitary eigenvalue problem. Only in the end the low rank matrix is reconstructed to retrieve the eigenvalues. Numerical experiments validate the approach, illustrate its reliability and speed. The algorithm is compared against other available methods and a proof of backward stability is provided. This research is joint work with David S.\ Watkins and Jared L.\ Aurentz from Washington State University, USA and Thomas Mach from the KU Leuven, Belgium. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Optimization & Numerical Analysis seminars: R. Vandebril
by Federico Poloni 01 Apr '15

01 Apr '15

Date: Giovedi' 9 Aprile Time: 11:00 Place: Aula Seminari, dipartimento di matematica Speaker: Raf Vandebril, Department of Computer Science, KU Leuven. Title: *The Companion QR* Abstract: In this lecture we will propose a new fast and stable manner of computing roots of polynomials. Roots of polynomials are typically computed by putting the coefficients of the polynomial in the companion matrix and then computing the eigenvalues of this matrix. Exploiting the available low-rank structure leads to an algorithm of quadratic instead of cubic complexity. Several low-cost algorithms have already been proposed. Either they fully exploit the low-rank properties of the involved matrices by representing them for instance via quasiseparable factors, or they write the companion matrix as the sum of a unitary plus low rank matrix and exploit this structure. In this lecture we will use the second manner. However, only a single QR step requires the use of the low rank part. After that we can put the low rank term aside and continue only with the unitary matrix, translating the problem thereby to a unitary eigenvalue problem. Only in the end the low rank matrix is reconstructed to retrieve the eigenvalues. Numerical experiments validate the approach, illustrate its reliability and speed. The algorithm is compared against other available methods and a proof of backward stability is provided. This research is joint work with David S.\ Watkins and Jared L.\ Aurentz from Washington State University, USA and Thomas Mach from the KU Leuven, Belgium. === Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 0

Fwd: [Dipartimento.di] Avviso Seminario: Leonardo Robol - Seminario di Esperienze di Programmazione
by Federico Poloni 31 Mar '15

31 Mar '15

Inoltro per altri possibili interessati. -------- Forwarded Message -------- Subject: [Dipartimento.di] Avviso Seminario: Leonardo Robol - Seminario di Esperienze di Programmazione Date: Mon, 30 Mar 2015 13:40:22 +0200 From: Gianna M. Del Corso <gianna.delcorso(a)unipi.it> To: dipartimento(a)di.unipi.it Nell'ambito del corso di Esperienze di Programmazione *Mercoledi' 1 Aprile 2015* - ore 11.00 Dip. di Informatica - Aula Seminari EST si terra' il seminario di LEONARDO ROBOL (Scuola Normale Superiore) INTRODUZIONE A JULIA Sarà presentato Julia, un linguaggio per il calcolo scientifico attualmente in fase di sviluppo. Verranno mostrate la sintassi e le principali differenze con i linguaggi attualmente più diffusi, come MATLAB e FORTRAN. Verrà fatta una panoramica delle principale caratteristiche della piattaforma, come i pacchetti aggiuntivi, il notebook IJulia, la gestione dei tipi e il raggiungimento di ottime performance (comparabili a quelle di C e FORTRAN) grazie all'utilizzo di LLVM come backend. Sarà infine presentato qualche esempio per mostrare le potenzialità di Julia per lo sviluppo di codice parallelo. L'interprete è open source e disponibile su http://julialang.org/. ======================== Riferimento: Gianna Del Corso Gianna M. Del Corso, PhD Dipartimento di Informatica, Universita' di Pisa Largo Pontecorvo, 3 56127 Pisa, Italy ph. +39-050-2213118 fax +39-050-2212726 Gianna M. Del Corso, PhD Dipartimento di Informatica, Universita' di Pisa Largo Pontecorvo, 3 56127 Pisa, Italy ph. +39-050-2213118 fax +39-050-2212726

1 0

Optimization & numerical analysis seminars: A. Frangioni
by Federico Poloni 19 Mar '15

19 Mar '15

In the next months we are holding a small cycle of joint seminars between the optimization and numerical analysis groups; their purpose is describing each other's research work, with an eye to possible new common areas of interest. Prof. Frangioni's seminar is the first one of the series. Time: Thursday 2015-Mar-19 (next week), h. 10:00 Place: Sala Seminari Ovest, Dipartimento di Informatica Speaker: Antonio Frangioni, Università di Pisa Title: *Iterative approaches for graph-structured linear systems arising in optimization* Abstract: When applying Interior Point (IP) algorithm to the solution of linear or convex separable optimization problems with strong graph structure, like the Min Cost Flow (MCF) problem, one has to repeatedly solve linear systems with strong graph structure. One of the two possible approach leads to systems with a M-matrix that is basically the weighted Laplacian of the underlying graph, where the weights (but not the graph) change at each iteration of the IP methods and typically become "very imbalanced" at the last ones. For this problem we have used two different techniques: Preconditioned Conjugate Gradient (PCG) methods, based on the idea of extracting a proper triangulated subgraph of the original graph which strictly contains a spanning tree. We have defined a new class of triangulated graphs, called Brother-Connected Trees (BCT), and discuss some fast heuristics for finding BCTs of "large" weight, starting from either the Kruskal algorithm or the Prim one. In addition we have studied multi-iterative techniques where the above PCG approach has been combined with specialized coarser-grid operators which preserve the graph structure of the projected matrix at the inner levels and smoothers, that have been completely characterized. Quite a lot of work still remains to do for understanding which is the best combination of preconditioning and coarsening at different stages of the overarching IP algorithm, implementing these techniques in an efficient solver, and extending them to "truly" interesting graph-structured problems like Multicommodity MCF ones. Everyone is welcome! -- --federico poloni Dipartimento di Informatica, Università di Pisa http://www.di.unipi.it/~fpoloni/ tel:+39-050-2213143

1 1

Seminario del Prof. J. Gondzio 20/3 a Firenze (RETTIFICA AULA)
by margherita porcelli 14 Mar '15

14 Mar '15

Carissimi, vi ricordo il Seminario del Prof. Jacek Gondzio (School of Mathematics, University of Edinburgh) di venerdì prossimo (20/3) a Firenze e, a rettifica di quanto comunicato in precedenza, preciso che il seminario si terrà presso il PLESSO DIDATTICO MORGAGNI (Aula 223). ********************************* Ven 20 Marzo, ore 14:00 Luogo: FIRENZE aula: Aula 223, PLESSO DIDATTICO MORGAGNI Viale Morgagni 40, Università di Firenze TITOLO: "Preconditioners for higher order methods in big data optimization" ABSTRACT: We address efficient preconditioning techniques for the second-order methods applied to solve various sparse approximation problems arising in big data optimization. The preconditioners cleverly exploit special features of such problems and cluster the spectrum of eigenvalues around one. The inexact Newton Conjugate Gradient method excels in these conditions. Numerical results of solving L1-regularization problems of unprecedented sizes will be presented. J. Gondzio joint work K. Fountoulakis ********************************* Cordiali saluti, Margherita Porcelli -- Margherita Porcelli Dipartimento di Matematica Università di Bologna Piazza di Porta S. Donato, 5 40127 - Bologna - Italy phone : +39-051-2094423 home page: http://web.math.unifi.it/users/porcelli ---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.

1 0

Seminari del Prof. Jacek Gondzio (University of Edinburgh)
by Margherita Porcelli 04 Mar '15

04 Mar '15

Segnalo 2 seminari che il Prof. Jacek Gondzio (School of Mathematics, University of Edinburgh) terrà nel corso della sua visita presso il Dipartimento di Ingegneria Industriale (UNIFI) e il Dipartimento di Matematica (UNIBO). ********************************* Ven 13 Marzo, ore 11:00 Luogo: BOLOGNA aula: Seminario I, Dipartimento di Matematica, Università di Bologna TITOLO: "Preconditioners for higher order methods in signal reconstruction" ABSTRACT: We address efficient preconditioning techniques for the inexact second-order methods applied to solve various sparse approximation problems arising in signal/image reconstruction. The preconditioners exploit two features of such problems: (i) sparsity of the solution, and (ii) near-orthogonality of the matrices involved. The latter originates from the restricted isometry properties frequently assumed in such applications. Spectral analysis of the preconditioners and their practical efficiency when solving linear systems in the Newton Conjugate Gradient method will be presented. If time permits then a few comments on the some other problems originating from the "Big Data" buzz will also be given. J. Gondzio joint work with I. Dassios, K. Fountoulakis and P. Zhlobich ********************************* Ven 20 Marzo, ore 14:00 Luogo: FIRENZE aula: Aula Tricerri, Dipartimento di Matematica e Informatica 'Ulisse Dini' Università di Firenze TITOLO: "Preconditioners for higher order methods in big data optimization" ABSTRACT: We address efficient preconditioning techniques for the second-order methods applied to solve various sparse approximation problems arising in big data optimization. The preconditioners cleverly exploit special features of such problems and cluster the spectrum of eigenvalues around one. The inexact Newton Conjugate Gradient method excels in these conditions. Numerical results of solving L1-regularization problems of unprecedented sizes will be presented. J. Gondzio joint work K. Fountoulakis ********************************* Cordiali saluti, Margherita Porcelli -- Margherita Porcelli Dipartimento di Matematica Università di Bologna Piazza di Porta S. Donato, 5 40127 - Bologna - Italy phone : +39-051-2094423 home page: http://web.math.unifi.it/users/porcelli

1 0

Reminder: Numerical Analysis Seminar, S. Simsek, 2014-09-23
by Federico Poloni 23 Sep '14

23 Sep '14

## Numerical Analysis Seminars, 2014-2015 Title: The matrix nearness problem for some special type of matrices associated with linear matrix equations Speaker: Sinem Simsek, Sakarya University and Kırklareli University, Turkey Date: Tuesday, 2014-Sep-23 Time: 14:00-15:00 Place: Sala Seminari, Dipartimento di Matematica Everyone is welcome! -- --federico

1 0

[Numpi.di] Numerical Analysis Seminar, S. Simsek, 2014-09-23
by Federico Poloni 18 Sep '14

18 Sep '14

## Numerical Analysis Seminars, 2014-2015 Title: The matrix nearness problem for some special type of matrices associated with linear matrix equations Speaker: Sinem Simsek, Sakarya University and Kırklareli University, Turkey Date: Tuesday, 2014-Sep-23 Time: 14:00-15:00 Place: Sala Seminari, Dipartimento di Matematica Everyone is welcome! -- --federico

1 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

NumPI