Struct AdmmSolver 

pub struct AdmmSolver {
    pub rho: f64,
    pub max_iter: usize,
    pub tol: f64,
}

Alternating Direction Method of Multipliers (ADMM) for problems of the form:

  minimise   f(x) + g(z)
  subject to x = z

Uses the augmented Lagrangian with penalty parameter rho.

Fields

rho: f64

Augmented Lagrangian penalty ρ > 0.

max_iter: usize

Maximum number of ADMM iterations.

tol: f64

Convergence tolerance on primal and dual residuals.

Implementations

impl AdmmSolver

pub fn new(rho: f64, max_iter: usize, tol: f64) -> Self

Create a new ADMM solver.

pub fn solve( &self, x_init: Vec<f64>, x_update: &dyn Fn(&[f64], &[f64], f64) -> Vec<f64>, z_update: &dyn Fn(&[f64], &[f64], f64) -> Vec<f64>, ) -> AdmmResult

Solve the consensus ADMM problem.

• x_update: solves argmin_x f(x) + (ρ/2)||x − z + u||².
• z_update: solves argmin_z g(z) + (ρ/2)||x − z + u||².