OPTEC Scientific Advisory Board Meeting
September 30, 2009, 9:00-18:00
CS building Celestijnenlaan 200A, ground floor
The yearly OPTEC Scientific Advisory Board (SAB) Meeting will be held
at Computer Science building - Celestijnenlaan 200A, ground floor.
Agenda
- 9:00- 9:30 Introduction and Overview of old OPTEC (seminar room) (Joos+Moritz)
- 9:30-10:30 coffee and poster session 1 (WP1+WP2) (Foyer)
- 10:30-11:00 two plenary "poster" presentations (Auditorium)
"Low-rank approximations by Riemannian optimization"
Bart Vandereycken
Division of Numerical Analysis and Applied Mathematics (NATW),
Department of Computer Science, K.U. Leuven
Abstract
Low-rank approximations are a common technique to reduce the
dimension of matrix problems. These problems include matrix equations,
like the
Lyapunov equation, and matrix completion, like the Netflix challenge.
In this work, we look at the low-rank approximation as an
optimization problem on a smooth manifold, namely the set of fixed-rank
PSD
matrices. By using the machinery of optimization on manifolds, we
employed standard trust-region methods for finding this low-rank
approximant. For certain large-scale PDEs, we have extended a
multi-level optimization strategy to manifolds and obtained
mesh-independent
convergence rates. In addition, we have analyzed the Riemannian
geometry of the fixed-rank matrices in two ways. By describing the set
as an embedding of the real matrices, it is more suited for efficient
optimization algorithms, while exploiting the homogeneous space
structure, we obtained a geodesically complete space.
"ACADO Toolkit - An Open-Source Framework for Automatic Control and Dynamic Optimization"
Boris Houska
Department of Electrical Engineering, K.U. Leuven
ACADO Toolkit is an open-source software environment and algorithm collection for automatic control and dynamic optimization that is currently developed within OPTEC. It provides a general framework for using a great variety of algorithms for direct optimal control, including model predictive control and state/parameter estimation. ACADO Toolkit is implemented as self-contained C++ code providing a very user-friendly syntax for setting up optimal control problems. The object-oriented design allows for convenient coupling of existing optimization packages and for extending it with user-written optimization routines. In this talk we discuss several of ACADO Toolkit's key features and demonstrate how it can be used to solve optimal control problems. Also the underlying mathematical theory is briefly explained.
- 11:00-12:00 coffee and poster session (WP3+WP4+WP5) (Foyer)
- 12:00-12:30 two plenary "poster" presentations (Auditorium)
"Multiple objective optimal control of (bio)chemical processes"
Filip Logist, Peter Van Erdeghem and Jan Van Impe
Chemical and Biochemical Process Technology and Control (BioTeC),
Department of Chemical Engineering, K.U. Leuven
In this talk we present a generic approach that allows the fast
generation of the set of Pareto optimal solutions for optimal control
problems with multiple and conflicting objective functions. The
presented approach combines novel and accurate scalarisation methods as
Normal Boundary Intersection and Normalised Normal Constraint with fast
direct optimal control techniques as multiple shooting. The approach is
illustrated for two (bio)chemical case-studies: (i) a fed-batch
bioreactor and (ii) a catalytic tubular reactor with periodic flow
reversals.
"Identifying customer profiles in power load times series using spectral clustering"
Carlos Alzate, Marcelo Espinoza, Bart De Moor and Johan Suykens
Department of Electrical Engineering, K.U. Leuven
An application of multiway spectral clustering with out-of-sample
extensions towards clustering time series is presented. The data
correspond to power load time series acquired from substations in the
Belgian grid for a period of 5 years. Spectral clustering methods are
a class of unsupervised learning algorithms where the solutions can be
obtained from the eigenvectors of a Laplacian matrix derived from the
data. A new formulation to spectral clustering is used to find
interpretable customer profiles underlying the power consumption load
time series. The main advantage of this approach is the extension of
the clustering model to out-of-sample points. The clustering model can
be trained, validated and tested in a learning framework working
directly with the data and without the use of pre-modeling steps. The
experimental results with real-life data demonstrate the applicability
of the multiway spectral clustering method compared to an existing
method pre-modeling the data based on periodic autoregressions (PAR).
- 12:30-13:45 lunch break with sandwiches, continuation of
posters session if desired
From now on in 5th floor, small seminar room:
- 13:45-14:00 Introduction into new OPTEC (JVI)
- 14:00-14:20 Intro Dept. Bouwkunde (Geert Degrande)
- 14:20-14:30 Intro into new Working Groups (MD)
- 14:30-14:40 WG1: MD
- 14:40-14:50 WG2: Johan S.
- 14:50-15:00 WG3: Filip Logist
- 15:00-15:30 coffee break
- 15:30-15:40 WG4: Geert L or Joris DS
- 15:40-15:50 WG5: Wim M
- 15:50-16:00 WG6: Johan M.
- 16:00-16:15 Integrative Projects - JSw
- 16:15-16:30 Closing and Questions to SAB
- 16:30-17:00 Closed Door Discussion of SAB
- 17:00-17:30 Communication of SAB findings: OLD OPTEC
- 17:30-18:00 Communication of SAB findings: NEW OPTEC, and closing
