Incorrect number or types of inputs or outputs for function vec.

cvx_begin

variable beam(K + 1, K) complex;

variable r_c(1, K);

variable R_p(1, K);

common_rate = sum(Group .* r_c);

rate = 0;

for i = 1 : K

upper_limit = R_p(i) + Group(i) .* r_c(i) - D_min(i);

rate = rate + (((theta * (1 - exp(- upper_limit * theta))) .* (upper_limit >= 0)) + upper_limit .* (upper_limit <= 0)) + D_min(i);

end

maximize(rate)

subject to

for i = 1 : K

g_private_i = Update_g_private_k(channel, beam, private_beta, alpha_private, Group, i, delt);

R_p(i) <= g_private_i;

0 <= R_p(i);

[R_c, ~] = Update_g_common_k(channel, beam, alpha_common, common_beta, Group, i, delt);

end

for j = 1 : K

0 <= r_c(j);

end

common_rate <= R_c;

% 功率约束

square_pos(norm(beam, 'fro')) <= P_max;

cvx_end

报错原因：

Incorrect number or types of inputs or outputs for function vec.

出错 cvx/sparsify>replcols (第 142 行)

cvx___.readonly( ndxs ) = vec( cvx_readlevel( bN ) );

出错 cvx/sparsify (第 61 行)

[ x, forms, repls ] = replcols( x, tt, 'full', forms, repls, isobj );

出错 cvx/exp (第 68 行)

xt = sparsify( xt, 'exponential' );

出错 Cvx_solver (第 49 行)

rate = rate + (((theta * (1 - exp(- upper_limit * theta))) .* (upper_limit >= 0)) + upper_limit .* (upper_limit <= 0)) + D_min(i);

Answer 1

[Ruchika Parag]

Hi, 麒麟,

我将用英语回答这个问题，因为我的母语不是中文.

The error you are encountering suggests that there is an issue with the 'exp' function in the CVX environment. This could be due to improper handling of the exponential function within the CVX optimization framework. Some steps and modifications you can try are as follows:

1. Vectorization and Element-wise Operations: Ensure that all operations involving vectors and matrices are properly vectorized and use element-wise operations where necessary.
2. CVX Compatibility: Ensure that the expressions inside the 'maximize' and 'subject to' blocks are compatible with CVX's constraints and objective function requirements.
3. Simplify Expressions: Break down complex expressions into simpler parts to make it easier to identify where the issue might be.

Here is a modified version of your code with these considerations:

cvx_begin
    variable beam(K + 1, K) complex;
    variable r_c(1, K);
    variable R_p(1, K);
    
    common_rate = sum(Group .* r_c);
    rate = 0;
    
    for i = 1 : K
        upper_limit = R_p(i) + Group(i) * r_c(i) - D_min(i);
        
        % Define a temporary variable for the exponential term
        exp_term = exp(-upper_limit * theta);
        
        % Ensure element-wise operations
        rate = rate + ((theta * (1 - exp_term)) * (upper_limit >= 0)) + upper_limit * (upper_limit <= 0) + D_min(i);
    end

    maximize(rate)
    subject to
        for i = 1 : K
            g_private_i = Update_g_private_k(channel, beam, private_beta, alpha_private, Group, i, delt);
            R_p(i) <= g_private_i;
            0 <= R_p(i);
            [R_c, ~] = Update_g_common_k(channel, beam, alpha_common, common_beta, Group, i, delt);
        end
        
        % Ensure non-negativity of r_c
        r_c >= 0;
        
        % Ensure common rate constraint
        common_rate <= R_c;
        
        % Power constraint
        square_pos(norm(beam, 'fro')) <= P_max;
cvx_end

By making these changes, you should be able to address the error and ensure that your optimization problem is correctly formulated for CVX. Hope this helps!

[Kyrin]

Thanks for the advice, I'm still having problems after modifying it this way.

Incorrect number or types of inputs or outputs for function vec.

出错 cvx/sparsify>replcols (第 142 行)
    cvx___.readonly( ndxs ) = vec( cvx_readlevel( bN ) );

出错 cvx/sparsify (第 61 行)
        [ x, forms, repls ] = replcols( x, tt, 'full', forms, repls, isobj );

出错 cvx/exp (第 68 行)
            xt = sparsify( xt, 'exponential' );

出错 Cvx_solver (第 51 行)
                exp_term = exp(-upper_limit * theta);