Approximating the set of efficient points efficiently
 
 
Description:  In multicriteria optimization, no point in the set of efficient points (the solution set) is able to represent the whole set: each efficient point is incomparable to each other one. Therefore, methods which can construct a representation of the whole set of efficient points are of particular interest. Since it is seldom the case that one is able to derive an explicit representation of the solution set, approximative methods have to be used. An efficient method for approximating the set of efficient points of a convex multiobjective problem is derived. The method is based on a warm-start technique for an interior-point algorithm for perturbed optimization problems. A complexity estimate shows that the efficiency of the method - measured in number of operations per point computed - increases with higher accuracy demands. Moreover, we present computational results for applications of the method to real-world portfolio optimization problems.
Area(s):
Date:  2001-10-30
Start Time:   14:00
Speaker:  Jörg Fliege (UniversitÀt Dortmund, Germany)
Place:  Room 5.5
Research Groups: -Numerical Analysis and Optimization
See more:   <Main>  
 
© Centre for Mathematics, University of Coimbra, funded by
Science and Technology Foundation
Powered by: rdOnWeb v1.4 | technical support