optimization-toolbox

General purpose optimization algorithms.

Algorithms for unconstrained optimization

• Steepest descent (Wolfe condition);
• Newton direction (possibly with Hessian modification);
• Trust region;
• Congjugate gradeint (Fletcher-Reeves);
• Quasi-Newton method - BFGS;
• Quasi-Newton method - SR1 with trust region;
• Newton method with conjugate gradient.

Algorithms for constrained optimization

PS: More algorithms are expected to be added.