We present a filter line-search algorithm that does not require inertia information about the linear system to ensure global convergence. The proposed approach performs curvature tests along the search step to ensure descent. This feature permits more modularity in the linear algebra, enabling the use of a wider range of iterative and decomposition strategies. We use the inertia-free approach in an interior-point framework and provide numerical evidence that this is as robust as inertia detection via symmetric indefinite factorizations and can lead to important reductions in solution time.
Article
View An Inertia-Free Filter Line-Search Algorithm for Large-Scale Nonlinear Programming