How can I translate the YALMIP expression to LMILAB expression?

3 次查看（过去 30 天）

Robert Ungi 2015-8-20

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/234796-how-can-i-translate-the-yalmip-expression-to-lmilab-expression

I have the following optimization problem implemented by the use of the YALMIP library:

    R11 = sdpvar(n_p, n_p);
    S11 = sdpvar(n_cl/2, n_cl/2);
    S12 = sdpvar(n_cl/2, n_cl/2);
    S22 = sdpvar(n_cl/2, n_cl/2);
    S = [S11 S12; S12' S22];
    gamma=sdpvar(1);
    % Positovity
    constr = [R11 > 0, S > 0, R11 - S11 > 0];
    % Open loop condition
    M = [R11*Ap'+Ap*R11         Bpw                             R11*Cpz';
        Bpw'             -gamma*eye(size(Bpw))                  Dpzw';
        Cpz*R11                 Dpzw                    -gamma*eye(size(Dpzw))]
    constr = [constr, M < 0];
    % Closed loop condition
    M = [S*Acl'+Acl*S         Bclw                             S*Cclz';
        Bclw'             -gamma*eye(size(Bclw',1))                  Dclzw';
        Cclz*S                 Dclzw                    -gamma*eye(size(Dclzw'))]

I would like to compare the solution by using the tools provided MATLAB (LMILAB), however I am stucked with the first step:

What is the equivalent expression for this:

    S = [S11 S12; S12' S22];
    gamma=sdpvar(1);
    % Positovity
    constr = [R11 > 0, S > 0, R11 - S11 > 0];

Should `S` be declared as a new lmi variable?

So far I started with this

setlmis([]);
R11=lmivar(1,[n_p, 1]);
S11=lmivar(1,[n_cl/2, 1]);
S12=lmivar(1,[n_cl/2, 1]);
S22=lmivar(1,[n_cl/2, 1]);
S=lmivar(1,[n_cl, 1]);
gamma=lmivar(1,[1, 0]);
lmiterm([-1 1 1 R11],1,1);                    
lmiterm([-2 1 1 S],1,1);                      
lmiterm([-3 1 1 R11],1,1);                      
lmiterm([-3 1 1 S11],1,-1);