Performance evaluation of an interior point filter line search method for constrained optimization

Abstract

Here we present a performance evaluation of three versions of a primal-dual interior point filter line search method for nonlinear programming. Each entry in the filter relies on three components, the feasibility, centrality and optimality, that are present in the first-order optimality conditions. The versions differ in a set of acceptance conditions that are used to consider a trial iterate to be acceptable to the filter. Performance profiles are used to compare the obtained numerical results in terms of the number of iterations and the number of the optimality measure evaluations.