Dear Mark,

Thanks for your answer. In my simulation, X is indeed a matrix. I follow your suggestion, however, I get a bad result, the status shows unbounded.

```
function [ a,objective2 ] = OptimalResourceAllocation( L_mu,h_M2U )
global SNR B Epsilon2 Epsilon1 K N delta P_static
cvx_begin
variables a(N,K) z(N,K)
expression E_mu(N,K)
for k=1:K
for n=1:N
E_mu(n,k)=(1./Epsilon1*(z(n,k)-a(n,k))*h_M2U(n,k)/SNR+Epsilon2*a(n,k)+P_static)*delta;
end
end
minimize sum(sum(E_mu))
subject to
for k=1:K
for n=1:N
{L_mu(n,k)./(delta*B)*log(2),a(N,K),z(N,K)} == exponential(1);
end
end
for n=1:N
sum(a(n,:))<=1;
end
a>=0.0001;
cvx_end
objective2=sum(sum(E_mu));
end
```

%%%%%%%%%%%%%%%

Successive approximation method to be employed.

Mosek will be called several times to refine the solution.

Original size: 2332 variables, 892 equality constraints

320 exponentials add 2560 variables, 1600 equality constraints

-----------------------------------------------------------------

Cones | Errors |

Mov/Act | Centering Exp cone Poly cone | Status

--------±--------------------------------±--------

0/ 0 | 0.000e+00 0.000e+00 0.000e+00 | Unbounded

-----------------------------------------------------------------

Status: Unbounded

Optimal value (cvx_optval): -Inf

```
a1 =
sparse matrix : 40×4
```

why this result incurs?