The Alternating Direction Methods of Multipliers (ADMM) is useful for optimization of goal functions with non-smooth terms, such as L1-norms.