Modern optimization methods and applications
Ekkehard W. Sachs(University of Trier, Germany)
Tuesday November 17, 11.00 a.m. CERFACS Conference Room
Abstract :
In this talk we consider constrained optimization problems for which efficient numerical solvers are developped. In many applications the variables can be classified into state and control variables. In this case often the constrained problem can be formulated as an unconstrained optimization problem at the expense of a highly nonlinear objective function.
We discuss various advantages and drawbacks of these two approaches. Emphasis is put on the use of iterative solvers for the subproblems which lead to inexact Newton and SQP-methods. Along these lines also interior point methods are discussed for the solution of optimization problems with inequality constraints.
Several applications are presented, in particular one from food sterilisation. This involves the solution of a discretized optimal control problem with partial differential equations which leads to large scale optimization problems with equality and inequality constraints.
Cerfacs' Conferences 1998 Home Page
