BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Automated parallel adjoints for model differentiation\, optimisati
 on and stability analysis - Patrick Farrell\, Imperial College London
DTSTART:20131017T103000Z
DTEND:20131017T113000Z
UID:TALK47509@talks.cam.ac.uk
CONTACT:Catherine Pearson
DESCRIPTION:The derivatives of PDE models are key ingredients in many impo
 rtant algorithms of computational science. They find applications in diver
 se areas such as sensitivity analysis\, PDE-constrained optimisation\, con
 tinuation and bifurcation analysis\, error estimation\, and generalised st
 ability theory.\n\nThese derivatives\, computed using the so-called tangen
 t linear and adjoint models\, have made an enormous impact in certain scie
 ntific fields (such as aeronautics\, meteorology\, and oceanography). Howe
 ver\, their use in other areas has been hampered by the great practical di
 fficulty of the derivation and implementation of tangent linear and adjoin
 t models. In his recent book\, Naumann (2011) describes the problem of the
  robust automated derivation of parallel tangent linear and adjoint models
  as "one of the great open problems in the field of high-performance scien
 tific computing”.\n\nIn this talk\, we present an elegant solution to th
 is problem for the common case where the original discrete forward model m
 ay be written in variational form\, and discuss some of its applications. 
LOCATION:Open Plan Area\, BP Institute\, Madingley Rise CB3 0EZ
END:VEVENT
END:VCALENDAR