One of the constraints evaluates to a FALSE LOGICAL variable

[studydog-pro]

Dear Johan, I encountered this error prompt while editing constraint conditions. May I know the reason? How can I solve this problem?What should I do with the inverse matrix of "PDN" that I have in mind?

[johanlofberg]

You're saying PDN*0 == 1, which trivially is impossible

https://yalmip.github.io/fixedconstraints

[studydog-pro]

Thank you for your help. But I set 'PDN_inv' to a variable of type sdpvar also shows this error. What should I do to get the inverse of a matrix, which is of type sdpvar?

[johanlofberg]

No it doesn't

>> PDN = sdpvar(2,2,3,'full');
>> PDN_inv = sdpvar(2,2,3,'full');
>> CS = [];for i = 1:3;CS = [CS, PDN_inv(:,:,i)*PDN(:,:,i) == eye(2)];end;CS
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
|   ID|                           Constraint|   Coefficient range|
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
|   #1|   Equality constraint (bilinear) 2x2|              1 to 1|
|   #2|   Equality constraint (bilinear) 2x2|              1 to 1|
|   #3|   Equality constraint (bilinear) 2x2|              1 to 1|
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

There is no good way to model inverse, unless you can perform a clever variable changes, congruence transformations, perform smart Schur complements etc

[studydog-pro]

Thanks for your patience in replying, I'll try other methods or change my model, thanks again!