Interior-Point Filter Line-Search

Abstract

AbstractIn this chapter, a combination called IPOPT of the interior-point algorithm with the filter line-search for solving large-scale nonlinear optimization problems, proposed by Wächter and Biegler (2005a, b), is presented. As we know, to allow convergence from poor starting points and to enforce progress to the solution, the interior point methods both in the trust-region and in the line-search frameworks with exact penalty merit function were developed. For example, KNITRO uses the l1 exact penalty function (Byrd, Gilbert, & Nocedal, 2000; Byrd, Hribar, & Nocedal, 1999).