On Friday 12 November 2010, the Centrum Wiskunde & Informatica (CWI) is organizing the symposium ‘Large-Scale and Uncertain Optimization’. Keynote speaker is Aharon Ben-Tal, head of the Minerva Optimization Center and professor at the Technion-Israel Institute of Technology. Currently he is holding a position at CWI as ‘Distinguished Scientist’. Ben-Tal has developed mathematical methods and computational algorithms for the solution of optimization problems that are, amongst others, used in medical applications and civil and aerospace engineering.
Optimization problems stem from numerous applications, most of which are affected by uncertainty: data are not accurate due to measurement errors, or the data being unavailable at the time decisions have to be made.
Radiotherapists, for example, are daily guided by optimization techniques in the planning of a radiation treatment for cancer. A treatment plan contains many elements of uncertainty. One crucial uncertainty is the actual size and location of the tumor which change over time due to movements or change in the breathing pattern of the patient. A relatively new methodology called Robust Optimization helps to design treatment programs which take into account a-priori the above uncertainties, thus ensuring a safe and effective radiation therapy.
Optimization techniques are used successfully in medical imaging. An example is the reconstruction of clinically acceptable images obtained from a PET (Positron Emission Tomography) scan. PET is a non evasive diagnostic method for locating cancerous tumors. The challenge here is to be able to locate the tumors when they are still small. The optimization problem associated with this problem is of a very large scale (up to several millions of variables) and special algorithms have to be devised to solve it in reasonable time and accuracy.
Another application area of large-scale optimization is in aerospace engineering. For instance mathematical techniques were used for the construction of a certain part in the wings of the new Airbus 380. A new design concept called FMO (Free Material Optimization) was employed here, which contrary to conventional techniques does not assume a given material from which to build the parts, but rather treats the material properties themselves as design variables. Using the FMO approach, combined with powerful optimization algorithms, a saving of 500 kg in the weight of each wing of the Airbus 380 aircraft was achieved.
Other speakers at the symposium are Marco Campi (University of Brescia), Dick den Hertog (Universiteit van Tilburg) and Arkadi Nemirovski (Georgia Institute of Technology).
The full programme with abstracts and registration form:
http://homepages.cwi.nl/~monique/Optimization-Symposium-Day/.
Picture: CT-PET scan, by J.C. Lopez
