Advances in Linear Algebra and Huge-Scale Optimization

1-2 July 2019

ICMS, The Bayes Centre, 47 Potterrow, Edinburgh EH8 9BT

Scientific Organisers

  • John Pearson, University of Edinburgh
  • Jacek Gondzio, University of Edinburgh
  • Luca Bergamaschi, Università degli Studi di Padova
  • Ángeles Martínez, Università degli Studi di Padova


It is well-established that bespoke numerical linear algebra techniques, including the design of direct and iterative methods, form an essential component of solvers for huge-scale problems in the fields of optimization, PDEs and their applications, and many others. In this workshop we focus on the fast and robust solution of matrix systems arising from a range of mathematical applications, and new challenges relating to modern optimization problems.

The meeting will consist of presentations from a number of leading researchers in optimization and numerical linear algebra, aiming to encourage an exchange of ideas to further advance the state-of-the-art in these fields.

Invited speakers

  • Luca Bergamaschi, Università degli Studi di Padova
  • Jordi Castro, Universitat Politècnica de Catalunya
  • Daniela di Serafino, Università degli Studi della Campania
  • Anders Forsgren, KTH Royal Institute of Technology
  • Daniel Loghin, University of Birmingham
  • Benedetta Morini, Università di Firenze
  • Margherita Porcelli (Università di Bologna
  • Spyros Pougkakiotis, University of Edinburgh
  • Miroslav Rozložník, Czech Academy of Sciences
  • David Silvester, University of Manchester
  • Jemima Tabeart, University of Reading
  • Arrangements

    Registration is now closed.

    Invited speakers accommodtion will be arranged for them by ICMS. Other participants should make their own arrangements. Please visit this page for some suggestions.

    ICMS does NOT use third parties to arrange accommodation. If you are approached by a third party (e.g. Business Travel Management) asking for booking details, please ignore. If you have any concerns please contact ICMS.

    Programme:

    The abstracts can be found here

    Monday 1 July


    10:30

    Registration  

    10:50

    Welcome

    11:00

    David Silvester (University of Manchester) Bespoke linear algebra for PDEs with random coefficients

    11:45

    Daniela di Serafino (Università della Campania "Luigi Vanvitelli") On constraint-preconditioned Krylov solvers for saddle-point linear systems

    12:30

    Lunch (not provided)

    14:30

    Anders Forsgren (KTH, Stockholm) On solving symmetric systems of linear equations arising in optimization

    15:10

    Tea and coffee

    15:30

    Luca Bergamaschi (Università di Padova) Efficient preconditioning of the normal equations for large LPs

    16:15

    Margherita Porcelli (Università di Bologna)  A new interior point approach for low-rank semidefinite programs

    17:00

    End of day one  

    19:00

    Dinner for invited speakers

    Tuesday 2 July


    09:30

     Jordi Castro (Universitat Politècnica de Catalunya)  A specialized interior-point algorithm for huge minimum convex cost flows in bipartite networks

    10:15

    Spyros Pougkakiotis (University of Edinburgh) Fast solution methods for convex fractional differential equation optimization

    11:00

    Tea and coffee

    11:30

    Miroslav Rozložník (Czech Academy of Sciences) Numerical behavior of saddle-point solvers

    12:15

    Jemima Tabeart (University of Reading/Imperial College London) Accounting for correlated errors in data assimilation: using linear algebra to improve computational efficiency

    13:00

    Lunch (not provided)

    14:15

    Daniel Loghin (University of Birmingham) Boundary preconditioning

    15:00

    Benedetta Morini (Università di Firenze)  Subsampled trust-region methods for finite-sum minimization

    15:45

    Workshop ends