Constraints with summation

[sb1368]

Hi everybody,

I have the following constraint in my optimization model:

sp(i,c)-s(i,c)+sum(j,sub(j,i,c))= 0 for all i and c

sp,s, and sub are decision variables.

How can I write this constraint using YALMIP? I examine two following codes to know the answer of this question but I get errors (obj is not important):

clc
clear

I=40; J=40; C=15;
s=sdpvar(I,C);
sp=sdpvar(I,C);
sub=sdpvar(J,I,C);

sumsub=sum(sub,1)
sumsubp(:,:)=sumsub(1,:,:)

sizes=size(s)
sizesend=size(sp)
sizesumsubp=size(sumsubp)

con=[sp-s+sumsubp==0]
obj=sum(sp)-sum(sumsubp);
sett=sdpsettings('solver','sedumi');
optimize(con,obj,sett);
Cost=value(obj)

Error using + (line 35)
Cannot add SDPVAR object and NDSDPVAR object

If I replace the above con with the below (By multiplying -1 ),

con=[-sp+s-sumsubp==0]

then the error is:
Error using |
NaN's cannot be converted to logicals.

Error in lmi/categorizeproblem>multipletermsInEquality (line 509)
p = any(sum(B | B,2)-(B(:,1) == 0) > 1);

Error in lmi/categorizeproblem (line 163)
if multipletermsInEquality(Fi)

Error in compileinterfacedata (line 298)
[ProblemClass,integer_variables,binary_variables,parametric_variables,uncertain_variables,semicont_variables,quad_info]
= categorizeproblem(F,logdetStruct,h,options.relax,parametric,evaluation_based,F_vars,exponential_cone);

Error in solvesdp (line 301)
[interfacedata,recoverdata,solver,diagnostic,F,Fremoved,ForiginalQuadratics] =
compileinterfacedata(F,[],logdetStruct,h,options,0,solving_parametric,solvers);

Error in solvesdp_multiple (line 27)
model = solvesdp(varargin{:});

Error in solvesdp (line 22)
diagnostic = solvesdp_multiple(varargin{:});

Error in optimize (line 31)
[varargout{1:nargout}] = solvesdp(varargin{:});

Error in Untitled7 (line 19)
optimize(con,obj,sett);

What is the best way to consider the mentioned constraint in an optimization modeling?

[johanlofberg]

Dimensions/class don't match. You probably want

sumsubp = squeeze(sumsub(1,:,:))

or

sumsubp = squeeze(sumsub)

[sb1368]

Many thanks for your help. The frist one works for me.