Dimensions of matrices being concatenated are not consistent ????

fati-a

unread,

Apr 5, 2018, 5:06:17 AM4/5/18

to YALMIP

Hello Johan

I have a problem about Dimensions of matrices, this problem is very strange for me because dimensions of parameter and controller are the same but still, there is this error :

Error using vertcat

Dimensions of matrices being concatenated are not consistent.

Error in eliminatevariables (line 249)

involved = [involved;find(m ~= fix(m) | m < 0)];

Error in optimizer/subsref (line 122)

[self.model,keptvariables,infeasible] =

eliminatevariables(self.model,self.parameters,thisData(:));

Error in cc_mpc2 (line 275)

[sol,err] = controller{{P.x_k,alpha_tot(24*20+k:24*20+k+P.Hp-1,:),...

could you please help me for solving this problem?

Thanks

cc_mpc2.m

forcastdemand.mat

Data3tanks.mat

eqcon_elimination.m

Dmodel.mat

Johan Löfberg

unread,

Apr 5, 2018, 5:46:32 AM4/5/18

to YALMIP

The optimizer does not like models which have signomial parameterizations in models that are intended to be solved with a simple QP solver. Hence, your A matrix should not be parameterized in x, but you should define new variables beta and let these be parameters, and compute these this from x before you call the controller object. That will also lead to much less overhead in every controller call, since a lot of symbolic processing can be removed

fati-a

unread,

Apr 5, 2018, 6:28:13 AM4/5/18

to YALMIP

Dear Johan

Thank you for quick answer.

I change the matrix A of my model and now the model is not signomial but still, there is the same problem .

Error using vertcat

Dimensions of matrices being concatenated are not consistent.

Error in cc_mpc2 (line 282)

[sol,err] = controller{{P.x_k,...

Thanks

cc_mpc2.m

Johan Löfberg

unread,

Apr 5, 2018, 6:45:35 AM4/5/18

to YALMIP

That's because the following code fails due to your commenting

>> {P.x_k,
%                                    alpha_tot(24*20+k:24*20+k+P.Hp-1,:),...
                                     alpha_tot(k:k+P.Hp-1,:),...
                                     D{mod(P.Hp+k-1,P.Hp)+1},...
                                     Dd{mod(P.Hp+k-1,P.Hp)+1},...
                                     bs{mod(P.Hp+k-1,P.Hp)+1},...
                                     devX{mod(P.Hp+k-1,P.Hp)+1}

                                     }
Error using vertcat
Dimensions of matrices being concatenated are not consistent.

...and once you've fixed that it will still fail as the dimensions in the parameter list is not consistent with the data that you send

Reply all

Reply to author

Forward