Non-Convex Solver
Non-Convex Solver
In this work, I developed and implemented a new solver for non-convex optimization problems, an infeasible-start primal-dual augmented-Lagrangian interior-point solver. It implements key algorithms employed by Ipopt to handle inequality constraint and a novel augmented-Lagrangian formulation and update to handle equality constraints. The primary advantage of this solver is its ability to natively handle complementarity constraints.
The solver can currently solve 41/41 benchmark problems (small size) from the CUTEst problem set and has been used to solve difficult contact-implicit trajectory optimization problems. The Julia implementation is available on Github.