Constrained Nonlinear Programming

Gianni Di Pillo and Laura Palagi

ABSTRACT

We consider the basic approaches for the solution of constrained nonlinear programming problems. The first approach is based on the use of unconstrained minimization techniques applied to some merit function, either in sequential penalty methods or in exact penalty methods. The second approach is based on the sequential solution of quadratic programming subproblems. The third approach tries to maintain feasibility of tentative solutions, by employing feasible descent directions. Far from being exhaustive, this article aims to give the main ideas and tools of practical use, and to suggest how to go deeper into more technical and/or more recent results.