development
and empirical comparative tests of optimization codes
nonlinear
programming algorithms, especially SQP methods
least
squares optimization by SQP-type methods
discretization and parameter
estimation
of dynamical systems (ODEs, DAEs, PDEs, PDAEs)
automatic
differentiation
topology
and free material optimization
parameter selection for support vector machines
mixed-integer nonlinear and nonconvex programming
IPM/SQP
methods for large-scale optimization
Industrial cooperations:
CAD-FEM, Grafing
Epcos, Munich
Boehringer
Ingelheim Pharma, Biberach
inuTech, Nurenberg
EADS, Manching
Astrium, Munich
TOMLAB
Optimization, Västerås
DYNARDO, Weimar
Shell, Rijswijk
ESTECO, Trieste
BASF, Ludwigshafen
... and others
How to reach me:
Klaus Schittkowski
Siedlerstr. 3
95488 Eckersdorf,
Germany
(+49)921
32887 klaus@schittkowski.de
secretary:
Department of Computer Science
University of Bayreuth, Germany
95440 Bayreuth
c/o Claudia Lachmann
(+49)921 557751 claudia.Lachmann@uni-bayreuth.de
... member of the EU-NoE
... member of EU-NoE
... member of the EU-STRP
... honorary member of
... member of the International Doctorate Programme
... member of the Shell GameChanger
programme
... subcontractor of the
FP7 Collaborative Project
This page was last updated on
11/02/11.
... click to see full picture
new:
Version 2.0 of EASY-OPTExpress
with continuous and integer/boolean variables, convexity not required
new:
NLPIP 1.0 solves large-scale nonlinear
optimization problems by a new SQP/IPM method (up to 5,000,000 variables)
new:
MISQP 6.2 solves 100 MINLP
test problems with less than 500 function calls including those needed for
derivative approximations
new:
A collection of 100 test problems for
nonlinear mixed-integer programming (with Fortran codes)
new:
NLPQLP 3.1 solves 95 % of our
test problems with only one correct digit in
the partial derivatives, termination at 10-7.
new: Version 5.1 of EASY-FITModelDesign
coming with priority levels, time and concentration depending fitting
criteria, etc.
new:
NLPQLB solves optimization problems
with 200,000,000 nonlinear constraints!
new:
Version 2.0 of
