4 - Smoot Unconstrained Optimization
Completion requirements
The simplest unconstrained optimization methods: gradient-type approaches. "Exact" line search. fixed stepsize or inexact Armijo-Wolfe-type Line Search. Dilation-type methods: Newton, Quasi-Newton, Limited-Memory Quasi-Newton. Deflection-type methods: Nonlinear Conjugate Gradient, Heavy Ball, a fleeting glimpse to fast gradient methods.
Click 4-smooth unconstrained optimization.pdf link to view the file.