Error using matlab.gra​phics.axis​.Axes/set Value

% Clear Old Variables

yalmip('clear'); C1=[]; C2=[]; C3=[]; C4=[];

Z0=sdpvar(nx); Z1=sdpvar(nu,nx); for ii=1:nz mu(ii)=sdpvar(1); end

C1=[Z0 >= 0, mu>=0];

C2=[[Z0-Gd*Sigw*Gd' (Ad*Z0-Bd*Z1); (Ad*Z0-Bd*Z1)' Z0] >= 0];

Izz=eye(nz);

for ii=1:nz rho_i=Izz(ii,:);

C3=[C3,[mu(ii) rho_i*(Dx*Z0-Du*Z1);(Dx*Z0-Du*Z1)'*rho_i' Z0]>=0];

end

% muj < zeta_maxj

for ii=1:nz C4=[C4, mu(ii) <= zeta_max(ii)]; end

Constraints=[C1,C2,C3,C4];

options = sdpsettings('verbose',0,'solver','mosek');

sol = optimize(Constraints,[],options);

if sol.problem == 0 % problem is feasible

Z0feas=value(Z0), Z1feas=value(Z1),

for ii=1:nz mufeas(ii)=value(mu(ii)); end

ZZfeas=(Dx*Z0feas-Du*Z1feas)inv(Z0feas)(Dx*Z0feas-Du*Z1feas)'

LLfeas=Z1feas/Z0feas, eig(Ad-Bd*LLfeas)

Sigx=dlyap(Ad-Bd*LLfeas,Gd*Sigw*Gd')

Sigz=(Dx-Du*LLfeas)Sigx(Dx-Du*LLfeas)'

qlabels=char('Mass Position (m)','Manipulated Variable (N)');

q_ossop=zeros(1,nz);qstar=zeros(1,nz); sigzmax=sqrt(zeta_max);

qmin=-sigzmax;qmax=sigzmax; SigmaZ=Sigz; alpha=1;

qqmin=1.0e+07 * [-3.8788 0 -0.0043]';qqmax=-qqmin;

xnum=2; ynum=1;

EDORplot(xnum,ynum,1,'y',q_ossop,qstar,qmin,qmax,SigmaZ,alpha^2,qlabels)

axis([qqmin(xnum) qqmax(xnum) qqmin(ynum) qqmax(ynum)]); box on;

set(gca, 'XTick', [-0.4:0.1:0.4]); set(gca, 'YTick', [-0.6 :0.1:0.6]);

set(gca,'FontName','Times New Roman','Fontsize',18)

elseif sol.problem == 1 % problem is infeasible

display('Infeasible Problem');

else

display('Hmm, something went wrong!'); sol.info; pause

end

[Walter Roberson]

Which line is the error occurring on? What are class() and size() of all the variables in that line?

[Harleen Kaur]

Hello,

It's right below EDOR plot :axis([qqmin(xnum) qqmax(xnum) qqmin(ynum) qqmax(ynum)]); box on;

A:6*6; B,G=1*6; Dx=2*6; Du=1*2; Sigx, sigz are calculated.

Answer 1

[Stephane Dauvillier]

Hi, please provide the line and the error message, it's really hard otherwise to provide a suitable answer.

On the following line for instance

 ZZfeas=(Dx*Z0feas-Du*Z1feas)inv(Z0feas)(Dx*Z0feas-Du*Z1feas)'

There is a missing times sign (*) before the inv function

ZZfeas=(Dx*Z0feas-Du*Z1feas)*inv(Z0feas)*(Dx*Z0feas-Du*Z1feas)'

By the way, the function yalmip doesn't exist (at least on the newest release).

If I may, I strongly encourage you to avoid more than one statement per line: it's unreadable.

For instance write

for ii=1:nz
    mu(ii)=sdpvar(1);
end

And not this

for ii=1:nz mu(ii)=sdpvar(1); end

Answer 2

[Walter Roberson]

Your qqmin sets the second component to 0. qqmax is the negative of that which is also 0. You then try to use that min and max as the x axes limits, but MATLAB complains because they are both 0. The upper bound for x must be different than the lower bound and neither can be nan. The upper can be inf but the lower cannot be -inf