Code:
variable W(N, N, K_max,M) complex hermitian semidefinite
for m = 1:M
for km = 1:K_max
[U, D] = eig(W_km_old{m}{km});
[lambda_max_val, idx] = max(diag(D));
u_max = U(:, idx); % 最大特征值对应的特征向量
W_lb{m}{km} = (lambda_max_val) + (trace((u_max * u_max’)’ * (W(:,:,km) - W_km_old{m}{km})));
end
end
obj = 0;
for m = 1:M
for km = 1:K_max
obj = obj + norm_nuc(W(:,:,km)) - (W_lb{m}{km}) ; % W_lb 你事先计算好
end
end
minimize( T + (1/(2*rho)) * obj )
There was an error when solving for W in CVX. How can I fix this issue?