My optimization problem is as follows:
variable NPAm(13,13,2) semidefinite (two 13*13 real PSD matrix)
objective function = NPAm(1,2,1)+NPAm(1,1,2)-NPAm(1,2,2) (a linear combination of some terms from variable NPAm)
all constraints are equalities:
The first part of these constraints states that some terms in NPAm(:,:,1) and NPAm(:,:,1) are equal
e.g. NPAm(1,2,1)==NPAm(2,2,1), NPAm(1,3,2)==NPAm(3,3,2), e.t.c.
The second part states that some terms on the same position of both NPAm(:,:,1) and NPAm(:,:,1) sum to a fixed real positive vector, e.g.
NPA(1,1,1)+NPA(1,1,2)==1;
NPA(1,2,1)+NPA(1,2,2)==0.5;
NPA(1,3,1)+NPA(1,3,2)==0.5;
NPA(1,4,1)+NPA(1,4,2)==0.5;
NPA(1,5,1)+NPA(1,5,2)==0.5;
…
The optimization is to maximise the objective function.
Now I only want to know the dual solutions associated to the second part of equality constraints, i.e. those sum constraints. So what can I do to check the dual solutions?