s.problem can be other number,for example,s.problem=-2 ?
if I use (s.problem == 1) | ((s.problem == 4) & (min(real(res)<-1e-3))) to judge a problem is infeasible(in my code,s.problem always equal to zero),that means this problem is always feasible?
But I don't think this problem is always feasible.So do you kown other ways to judge a problem is feasible or not in my matlab code?
Generally speaking ,Yalmip can solve this problem correctly?

Thanks ,lofberg !

users.isy.liu.se/johanl/yalmip/pmwiki.ph...Commands.Yalmiperror

Well, of course it is not always feasible, since you run a bisection.

s.problem always 0?? That does not happen on my machine when I run the code (I hope you are using SeDuMi or similiar. You cannot use LMILAB for bisection since LMILAB does not return infeasibility messages to YALMIP). U goes from around 1.0 and L from roughly 0.5, and they converge to a number in between after a couple of iterations, with error codes either 0 or 1 (i.e. numerical issues never occur)

Yes,you are right,I found I used LMILAB for bisection.
I have already installed SeDuMi_1_3.But I found a serious problem that the calculation using SeDuMi_1_3 is not accurate.

for example: a feasibility of the following code is
X=[1.0329 0.4647 -0.23583;
0.4647 0.77896 -0.050684;
-0.23583 -0.050684 1.4336];
----------------------------
code：
------------------------------

A=[-2 -2 -1;-3 -1 -1;1 0 -4];
B=[-1 0;0 -1;-1 -1];
Q=[-2 1 -2;1 -2 -4;-2 -4 -2];
R=eye(2);
X=sdpvar(3,3);
Co=[A'*X+X*A+Q,X*B;B'*X ,-R];
F=set(X>0)+set(Co<0);
s=solvesdp(F);
X=double(X);
---------------------------------------

But after running ,the matlab give a wrong answer:

X =

0.6195 0.1460 -0.1987
0.1460 0.5037 -0.2447
-0.1987 -0.2447 0.8561

------------------------------------------

why Semidu can't give a accurate result?

What do you mean with wrong answer? You have no objective and there is no reason to suspect a unique feasible solution

Thanks lofberg,I understand.