Why Cannot perform the operation: {real affine} ./ {real affine}

Nonconvex
1 
 cvx_begin
           variables Ps R
           maximize ((1-t).*R./Cmax - t.*(Ps+Pc)./Emax)
           subject to            
              Ps>0
              Ps<=5
               R<=log(1+s.*r.*Ps.*hsr./(1-s.*r.*hrr).*hrd./N0)/log(2)
               R<=log(1+(1-r).*Ps.*hsr.*(1-s.*r.*hrr)./((1-r).*s.*r.*Ps.*hsr.*hrr + N0.*(1-s.*r.*hrr)))/log(2)
               R>=0.1
               Ps.*(s.*r.*hsr.*hrp./(1-s.*r.*hrr) + hsp)<=4
        cvx_end

but Cannot perform the operation: {real affine} ./ {real affine}

2

Standard linear-fractional programming does not apply given the nonlinear constraints.

Have you proven this is a convex optimization problem, or can be transformed into a convex optimization problem? I will assume is is non-convex, unless you prove otherwise.

3

Thank you. I have solved it.

4

Perhaps you can tell us how you solved it.

5

R<=log(1+(1-r).*Ps.hsr.(1-s.*r.*hrr)./((1-r).*s.*r.*Ps.*hsr.hrr + N0.(1-s.*r.*hrr)))/log(2)
The transformation takes the SCA approach。