源代碼出自boyd
https://web.stanford.edu/~boyd/papers/admm/basis_pursuit/basis_pursuit.html
給出詳細備註
function [z, history] = basis_pursuit(A, b, rho, alpha)
% 解決如下 ADMM問題:
% minimize ||x||_1
% subject to Ax = b
%其中返回的 history變量包含 目標值、原始殘差和對偶殘差,以及每次迭代時原始殘差和對偶殘差的容差(the objective %value, the primal and dual residual norms, and the tolerances for the primal and dual residual norms at each iteration.)
% rho is the augmented Lagrangian parameter.
% alpha is the over-relaxation parameter (typical values for alpha are between 1.0 and 1.8).
t_start = tic;
% Global constants and defaults
QUIET = 0;
MAX_ITER = 1000;
ABSTOL = 1e-4;
RELTOL = 1e-2;
%% Data preprocessing
[m n] = size(A);
%% ADMM solver
基本上是按paper中寫的順序來的
x = zeros(n,1);
z = zeros(n,1);
u = zeros(n,1);
if ~QUIET
fprintf('%3s\t%10s\t%10s\t%10s\t%10s\t%10s\n', 'iter', ...
'r norm', 'eps pri', 's norm', 'eps dual', 'objective');
end %不斷輸出狀態量
% precompute static variables for x-update (projection on to Ax=b)
AAt = A*A';
P = eye(n) - A' * (AAt \ A);
q = A' * (AAt \ b);
for k = 1:MAX_ITER %迭代
% x-update
x = P*(z - u) + q;
% z-update with relaxation
zold = z;
x_hat = alpha*x + (1 - alpha)*zold;
%x這兒做了一個over-relaxation
z = shrinkage(x_hat + u, 1/rho);
%shrinkage是一個soft thresholding operator,如下所示:
u = u + (x_hat - z);
% diagnostics, reporting, termination checks
history.objval(k) =norm(x,1); %目標函數
history.r_norm(k) = norm(x - z);
history.s_norm(k) = norm(-rho*(z - zold));
history.eps_pri(k) = sqrt(n)*ABSTOL + RELTOL*max(norm(x), norm(-z));
history.eps_dual(k)= sqrt(n)*ABSTOL + RELTOL*norm(rho*u);
%這邊是如下的一種終止條件,
if ~QUIET
fprintf('%3d\t%10.4f\t%10.4f\t%10.4f\t%10.4f\t%10.2f\n', k, ...
history.r_norm(k), history.eps_pri(k), ...
history.s_norm(k), history.eps_dual(k), history.objval(k));
end
if (history.r_norm(k) < history.eps_pri(k) && ...
history.s_norm(k) < history.eps_dual(k))
break;
end
end
if ~QUIET
toc(t_start);
end
end
function y = shrinkage(a, kappa)
y = max(0, a-kappa) - max(0, -a-kappa);
end