Homework 5 for Convex Optimization in Autumn 2018 lectured by Prof. Zaiwen Wen. Submitted Version.
This repository implements various algorithms to solve LASSO problem via Matlab, including:
- Projection gradient method by reformulating the primal problem as a quadratic program with box constraints
- Subgradient method for the primal problem
- Gradient method for the smoothed primal problem
- Fast gradient method for the smoothed primal problem
- Proximal gradient method for the primal problem
- Fast proximal gradient method for the primal problem
- Augmented Lagrangian method for the dual problem
- Alternating direction method of multipliers for the dual problem
- Alternating direction method of multipliers with linearization for the primal problem
- AdaGrad, Adam, RMSProp, Momentum using subgradient