Codes for "Nonlinear Optimization with Geometric Constraints" (Master's thesis)

This repository contains a C++ implementation of optimization algorithms for problems with geometric constraints.

The code supports two main problem types.

Quadratic programming problems

min 1/2 x^T Q x + q^T x s.t. x in X, Ax - b in C.

Here, X and C are handled through projection routines. The implemented solvers include projected-gradient subsolvers and augmented Lagrangian methods, mainly ALMsolver1 and ALMsolver4.

Nonlinear SQP problems

min phi(x) s.t. x in X, Ax - b in C.

The nonlinear problem is solved by a local SQP method. At each SQP iteration, a quadratic subproblem is constructed and solved using the available QP solvers.

More details in the README file.