Duality Bounds in Global Optimization
 
 
Description:  A survey is given of some astonishingly promising recent results in the use of (Lagrangian-)dual bounds in branch and bound methods for the global minimization of nonconvex functions over convex and nonconvex sets.In these methods,the calculation of good lower bounds is crucial. After a brief glance at a typical branch and bound scheme the main results are presented.These include a comparison with convex envelopes,convergence and various applications.From a practical viewpoint the main result is that for many important problem classes the dual bound is better then the uniformly best convex lower bound (convex envelope-) minimization ,and, moreover,it turns out to require only solving a single linear program.
Area(s):
Date:  2005-09-20
Start Time:   12:00
Speaker:  Reiner Horst (University of Trier,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