Why Status: Unbounded Optimal value (cvx_optval): -Inf

1

my code:

cvx_solver sedumi
cvx_begin
variable Bc(G,1)
minimize (absum(rel_entr(lamda_g+nballoctBc+guiyihuaxishu_col.(P-sum(Po)-Pc)./(Ug-1),fenzi_col.Pc+lamda_g+nballoctBc+guiyihuaxishu_col.(P-sum(Po)-Pc)./(Ug-1))/n…
-(lamda_g/n+guiyihuaxishu_col.(P-sum(Po)-Pc)./((Ug-1)n)).log(1-(fenzi_col.Pc).inv_pos(fenzi_col.Pc+lamda_g+nballoctBc+guiyihuaxishu_col.(P-sum(Po)-Pc)./(Ug-1))))/log(2)+ …
(c/2).(sum(Bc + Bo)-B )^2 + lamda2(sum(Bc)-B))
subject to
Bc>=B/G/9;
sum(Bc + Bo)-B <=0;
cvx_end

image
image860×710 16.5 KB

my optimation question: max the data rate
image
image1040×390 24 KB

2

If you have Mosek available, specify it as solver. Otherwise follow the directions at CVXQUAD: How to use CVXQUAD's Pade Approximant instead of CVX's unreliable Successive Approximation for GP mode, log, exp, entr, rel_entr, kl_div, log_det, det_rootn, exponential cone. CVXQUAD's Quantum (Matrix) Entropy & Matrix Log related functions .

And make sure all the non-zero input data is within s small number of orders of magnitude of 1. Change units if needed to accomplish this.

if the model is still reported as unbounded, follow the advice at
Debugging unbounded models - YALMIP , which also applies to CVX.