Maximize sum log2(inv_pos(x))

maximize(obj) 
subject to
for k=1:N
    y(k)>=c_bar(k)^2+d_bar(k)^2+2*c_bar(k)*(c(k)-c_bar(k))+2*d_bar(k).*(d(k)-d_bar(k))%: z{k};
end
%obj=0;
%y(k)가 문제네
for k=1:N
    obj=obj+log(1+T*inv_pos(a*y(k)+b))/log(2)
end
subject to
for f=1:M
    abs(phi(f))<=1
end
cvx_end

I’m solving this problem and trying to maximize sum rate capacity, and got the problem above. I wonder if I can modify the problem above for CVX

No. sum log2(inv_pos(x)) is convex, therefore can only be minimized in cvx.