# {convex} .^ {0.5}

(a^2+pow_pos(norm(A-B),2))^(1/2)
as
norm([a;norm(A-B)])

Disciplined convex programming error:
Cannot perform the operation norm( {mixed convex/constant},
2 )

Are you sure you do not mean

norm([a;A-B])

?

A-B is convex. a is constant. Then “Cannot perform the operation norm( {mixed convex/constant},
2 )”

Given you have

||f(x)||

then f must be an affine expression of the variables e.g. x. Most likely that is your mistake.

Thank you very much.

I have one more question to ask you.When I used the MoSEk solver, "Status: inaccurately /Unbounded Optimal Value (CVX_OPtVAL): +Inf
" would appear after several iterations.

## Calling Mosek 9.1.9: 10430 variables, 5151 equality constraints For improved efficiency, Mosek is solving the dual problem.

MOSEK Version 9.1.9 (Build date: 2019-11-21 11:34:40)
Platform: Windows/64-X86

MOSEK warning 57: A large value of 1.9e+09 has been specified in c for variable ‘’ (241).
MOSEK warning 57: A large value of 1.1e+09 has been specified in c for variable ‘’ (482).
MOSEK warning 57: A large value of 3.7e+08 has been specified in c for variable ‘’ (723).
MOSEK warning 57: A large value of 2.1e+09 has been specified in c for variable ‘’ (964).
MOSEK warning 57: A large value of 2.8e+09 has been specified in c for variable ‘’ (1205).
MOSEK warning 57: A large value of 1.8e+08 has been specified in c for variable ‘’ (1446).
MOSEK warning 57: A large value of 8.5e+08 has been specified in c for variable ‘’ (1687).
MOSEK warning 57: A large value of 1.8e+09 has been specified in c for variable ‘’ (1928).
Problem
Name :
Objective sense : min
Type : CONIC (conic optimization problem)
Constraints : 5151
Cones : 2516
Scalar variables : 10430
Matrix variables : 0
Integer variables : 0

Optimizer started.
Presolve started.
Linear dependency checker started.
Linear dependency checker terminated.
Eliminator started.
Freed constraints in eliminator : 2518
Eliminator terminated.
Eliminator started.
Freed constraints in eliminator : 0
Eliminator terminated.
Eliminator - tries : 2 time : 0.00
Lin. dep. - tries : 1 time : 0.02
Lin. dep. - number : 0
Presolve terminated. Time: 0.05
Problem
Name :
Objective sense : min
Type : CONIC (conic optimization problem)
Constraints : 5151
Cones : 2516
Scalar variables : 10430
Matrix variables : 0
Integer variables : 0

Optimizer - solved problem : the primal
Optimizer - Constraints : 1786
Optimizer - Cones : 2515
Optimizer - Scalar variables : 7902 conic : 7664
Optimizer - Semi-definite variables: 0 scalarized : 0
Factor - setup time : 0.01 dense det. time : 0.00
Factor - ML order time : 0.00 GP order time : 0.00
Factor - nonzeros before factor : 9.71e+04 after factor : 1.29e+05
Factor - dense dim. : 85 flops : 3.92e+07
ITE PFEAS DFEAS GFEAS PRSTATUS POBJ DOBJ MU TIME
0 2.4e+03 5.2e+06 1.5e+07 0.00e+00 -1.102051416e+10 -1.103596091e+10 1.0e+00 0.08
1 5.0e+02 1.1e+06 6.9e+06 -1.00e+00 -1.102049804e+10 -1.103594426e+10 2.1e-01 0.17
2 1.4e+02 3.0e+05 3.6e+06 -1.00e+00 -1.102043530e+10 -1.103587896e+10 5.7e-02 0.19
3 6.6e+01 1.4e+05 2.5e+06 -9.99e-01 -1.101982274e+10 -1.103525861e+10 2.7e-02 0.20
4 4.4e+01 9.3e+04 2.0e+06 -9.98e-01 -1.101751760e+10 -1.103294366e+10 1.8e-02 0.20
5 3.9e+01 8.2e+04 1.9e+06 -9.97e-01 -1.101516700e+10 -1.103059338e+10 1.6e-02 0.22
6 2.3e+01 5.0e+04 1.5e+06 -9.97e-01 -1.100606393e+10 -1.102147821e+10 9.6e-03 0.23
7 1.2e+01 2.6e+04 1.1e+06 -9.96e-01 -1.098394721e+10 -1.099933128e+10 5.0e-03 0.25
8 2.1e+00 4.4e+03 4.4e+05 -9.92e-01 -1.076479248e+10 -1.077986494e+10 8.6e-04 0.27
9 4.5e-01 9.6e+02 1.9e+05 -9.63e-01 -1.028231643e+10 -1.029669984e+10 1.9e-04 0.28
10 1.5e-01 3.3e+02 1.1e+05 -9.06e-01 -9.643779833e+09 -9.657245007e+09 6.3e-05 0.30
11 6.0e-02 1.3e+02 5.9e+04 -8.33e-01 -8.784402969e+09 -8.796619866e+09 2.4e-05 0.31
12 2.6e-02 5.4e+01 3.4e+04 -7.33e-01 -7.841014303e+09 -7.851830667e+09 1.0e-05 0.33
13 6.6e-03 1.4e+01 1.3e+04 -6.05e-01 -5.726811014e+09 -5.734416620e+09 2.7e-06 0.33
14 1.9e-03 4.0e+00 4.1e+03 -3.98e-01 -3.925056082e+09 -3.929717536e+09 7.6e-07 0.34
15 4.4e-04 9.3e-01 8.5e+02 -8.01e-02 -2.311147994e+09 -2.313056929e+09 1.8e-07 0.36
16 6.2e-05 1.3e-01 6.1e+01 4.84e-01 -1.406055710e+09 -1.406407353e+09 2.5e-08 0.38
17 9.6e-06 2.0e-02 3.8e+00 9.04e-01 -1.233861119e+09 -1.233917985e+09 3.9e-09 0.39
18 3.6e-07 7.7e-04 2.9e-02 9.78e-01 -1.202157238e+09 -1.202159416e+09 1.5e-10 0.41
19 1.4e-07 1.2e-04 1.8e-03 1.00e+00 -1.201094251e+09 -1.201094596e+09 2.3e-11 0.41
20 1.4e-07 1.2e-04 1.8e-03 1.00e+00 -1.201094251e+09 -1.201094596e+09 2.3e-11 0.45
21 1.4e-07 1.2e-04 1.8e-03 1.00e+00 -1.201094251e+09 -1.201094596e+09 2.3e-11 0.47
Optimizer terminated. Time: 0.53

Interior-point solution summary
Problem status : UNKNOWN
Solution status : UNKNOWN
Primal. obj: -1.2010942514e+09 nrm: 2e+07 Viol. con: 6e-01 var: 2e-05 cones: 1e-05
Dual. obj: -1.2010945965e+09 nrm: 3e+09 Viol. con: 0e+00 var: 1e+02 cones: 0e+00
Optimizer summary
Optimizer - time: 0.53
Interior-point - iterations : 22 time: 0.50
Basis identification - time: 0.00
Primal - iterations : 0 time: 0.00
Dual - iterations : 0 time: 0.00
Clean primal - iterations : 0 time: 0.00
Clean dual - iterations : 0 time: 0.00
Simplex - time: 0.00
Primal simplex - iterations : 0 time: 0.00
Dual simplex - iterations : 0 time: 0.00
Mixed integer - relaxations: 0 time: 0.00

Status: Inaccurate/Unbounded
Optimal value (cvx_optval): +Inf

Lo =

1.0e+03 *

``````1.7566
``````

-0.0000

Mosek provided a warning about large values in the input data. It appears the scaling of the input data is bad. Try to improve it.by changing units, so that non-zero input data is within s small number of orders of magnitude of 1.

I’ve given you an initial set of values that I can solve for.The problem above arises when we solve it the second time.

Ql = Qr;
t1l=t1f;
t2l=t2f;
l = 0;
tol = tolerance;
Lo = [];
for l = 1: 1000
Qlt = repelem(Ql,1,1,K);
J = H^2 + sum( (Qlt - w).^2 ,1);
J = reshape(J, [N,K] );
I = P.dB2dec(beta0)(alpha/2)log2(exp(1)) ./ (J . (dB2dec(sigma_2) * J .^(alpha/2) + P.dB2dec(beta0)));
A = log2(1 + (P.dB2dec(beta0)) ./ (dB2dec(sigma_2) * J .^(alpha/2)) );
cvx_clear
cvx_begin %quiet
cvx_solver mosek
variables Q(2,N)
variable lo(1)
variable t1(N)
variable t2(N)
expression LO(K)
expression Ef(1,N)
expression Ef2(1,N)
expression WPT(1,N)
expression WPT2(1,N)
maximize lo %%
subject to:
for k = 1 : K
for m=1:N
Rlb =B
delta_t
( A(m,k) - I(m,k) * pow_pos(norm( Q(:,m) - u(:,k)),2) + I(m,k) * pow_pos(norm( Ql(: , m ) - u(:,k)),2));
LO(k) = LO(k) + Xr(m,k) * Rlb;
end
LO(k)>=D(k); %(10b)

`````` end
sum(LO)>=lo;
for m = 2: N
norm(Q(:,m) - Q(:,m-1)) <= Dmax;%(1)
end
norm(Q(:,1) - Q(:,N)) <= Dmax;
Q(1,1) == Q(1,N);
Q(2,1) == Q(2,N);

for m=2:N
norm([H;Q(:,m)-Gr(:,m)])+(-rel_entr(1,t1l(m)))/afa+(t1(m)-t1l(m))/(afa*t1l(m))<=0;
end
for m=2:N
2*t2(m)+t1(m)^2+Ps^2+(t1l(m)+Ps)^2-2*(t1l(m)+Ps)*(t1(m)+Ps)<=0;
end
Ef(1,1)=0;
Ef2(1,1)=0;
WPT(1,1)=0;
WPT2(1,1)=0;
for m=2:N
Ef(1,m)=Ef(1,m-1)+0.5*M*delta_t*pow_pos(norm(Q(:,m) - Q(:,m-1),2),2)/delta_t^2;
WPT(1,m)=WPT(1,m-1)+delta_t*(c1*t2(m)+c2*t1(m)+c3);
0-E0-WPT(1,m)+Ef(1,m)<=0;
Ef2(1,m)=Ef2(1,m-1)+0.5*M*delta_t*((-pow_pos(norm(Ql(:,m)-Ql(:,m-1)),2)+2*(Ql(:,m)-Ql(:,m-1))'*(Q(:,m)-Q(:,m-1)))/(delta_t^2));
WPT2(1,m)=WPT2(1,m-1)+delta_t*(c1*t2(m)+c2*t1(m)+c3);
Ef2(1,m)-WPT2(1,m)>=0;
end
``````

cvx_end
%%
figure(1)
hold on
plot(Q(1,:),Q(2,:), ‘-.’);
hold on
Lo = [Lo;lo/B]
Ql = Q;
t1l=t1;
t2l=t2;
if (l >= 2) &&(Lo(l) - Lo(l-1)<0.01)
break;
end
end

Perhaps your overall iterative scheme is unstable, producing wilder and wilder results, which become inputs for the next iteration, eventually resulting in failure. You can search this forum for my comments on Successive Convex Approximation.