Ask for help to solve the problem

Nonconvex
1

Hello
My problem here is the code

When writing this code I encounter the following error

Undefined function or method ‘log2’ for input arguments of type ‘cvx’.
Code is written by me as follows:
clc
clear all
close all
%%
N = 1000; % number of bits or symbols
h =1/sqrt(2)[randn(1,N) + jrandn(1,N)]; % Rayleigh channel
H = 1/sqrt(2)[randn(1,N) + jrandn(1,N)];
%%
p1=2;
p2=10;
pc=1;
alpha=10;
NT1=2;
NT2=2;
NR=2;
B=180*(10^3);
F=3;
N0=-174;
sigma=FBN0;
i=ones(1,N);
%%
k21_0=randn(1,N);
k22_0=randn(1,N);
Q1=(p1./(sum(diag(h))).^2)(Hctranspose(H)hctranspose(h));
R1=@(k21,k22) log2(1+(p1*(sum(diag(h))).^2)/(sigma+(k21+k22).hctranspose(h)));
R12=@(k22) log2(abs(i+(sigma.*i+H.k22ctranspose(H)).^(-1).*Q1));

R2=@(k21,k22) log2(abs(i+inv(sigma).H.(k21+k22)*ctranspose(H)+inv(sigma)*Q1))-R12;

Pint=@(R1) ((p1.(norm(h)).^2)./(2.^((R1))-1))-sigma;
R=log2(1+(p1(norm(h))^2/sigma));

N=@(k21) log2(abs(i+k21.hctranspose(h)+(sigma*i+Q1).^(-1)));
D=@(k21) alpha.*sum(diag(k21))+pc;
q=N(k21_0)/D(k21_0);
%%
cvx_begin

variables k21

N=log2(abs(i+k21.hctranspose(h)+(sigma*i+Q1).^(-1)));
D=alpha.*sum(diag(k21))+pc;

EE=N-q.D;
minimize N
subject to
R(k21)>=R
R12>=R
trace(k21)<=p2
cvx_end
Please help me solve this problem
Thanks

2

log2 is not among the functions CVX supports; but of course, since log_2(x)=log(x)/log(2), you don’t need it.

That said, I doubt CVX can solve this problem. Please see Why isn't CVX accepting my model? READ THIS FIRST!

3

Hello
You can help me how I issue the following simulation

Thanks