On the min-max graph in finite and semi-infinite optimization

 

                                 Hubertus Th. Jongen

 

               RWTH Aachen University (D) and University Maastricht (NL)

 

               We consider finite dimensional smooth optimization problems with compact

               connected feasible set. A variable (= Riemannian) metric defines an  ascent and

              descent semi-flow. This gives rise to a bipartite digraph on the set of local

              minimizers and maximizers  (min-max graph). Active set strategy may cause a

             stable obstruction to the connectedness of the min-max graph. However, by means

             of an automatic constraint-adaptation of the metric the min-max graph becomes

             generically connected. In case of a single smooth inequality constraint

             We give an explicit formula of the metric adaptation. In case of finitely many

             constraints we propose a logarithmic pre-smoothing and for semi-infinite

            optimization we discuss a mollifier-pre-smoothing. 

               This is joint work with Oliver Stein.

 

                References:

1.     Jongen, H.Th., Stein, O.:  Constrained global optimization: adaptive gradient

      flows.  In: Frontiers in Global Optimization (Eds.: C.A. Floudas, P. Pardalos),

     Kluwer Academic Publishers, Boston, pp. 223-236 (2004).

2.     Floudas, C.A., Jongen, H.Th.:  Global optimization: local minima and transition

    points. Journal of Global Optimization, pp. 409-415 (2005)           

          3. Jongen, H.Th., Stein, O.:
              Smoothing by mollifiers. Part I: Semi-infinite optimization.
              Journal of Global Optimization (to appear)

          4.Jongen, H.Th., Stein, O.:
             Smoothing by mollifiers. Part II: Nonlinear optimization.
             Journal of Global Optimization (to appear)