From: Yaquan Xu on
Matt,


Thank you very much for your software. I can use your tool to solve my case more efficient.

the expanded matrix is used to solve the clique partition problem in quadratic way. The first matrix is just a test sample for the weighted graph with 4 nodes ( actually it is not a good test sample since it is a clique already)

The size of expanded matrix is based on the number of cliques. assuming if there is two cliques in the graph, then we have 8 binary variables (Xij, 1 if node i assigned to clique j, 0 otherwise) and the expanded matrix goes to 8 by 8. If there is three cliques, then the matrix goes to 12 by 12, etc.




"Matt J " <mattjacREMOVE(a)THISieee.spam> wrote in message <hnimgq$b8u$1(a)fred.mathworks.com>...
> "Yaquan Xu" <yaquanxu(a)gmail.com> wrote in message <hni6u4$pb2$1(a)fred.mathworks.com>...
> > Thank you so much. It really solved my problem!
>
> But you might want to explain a bit what you need this for.
> It's important to point out that one rarely needs to explicitly expand a matrix using
> kron(A,B), as you're doing. It's usually much more efficient to work with A and B separately, which is why I made this tool:
>
> http://www.mathworks.com/matlabcentral/fileexchange/25969-efficient-object-oriented-kronecker-product-manipulation
From: Yaquan Xu on
Matt,


Thank you very much for your software. I can use your tool to solve my case more efficient.

the expanded matrix is used to solve the clique partition problem in quadratic way. The first matrix is just a test sample for the weighted graph with 4 nodes ( actually it is not a good test sample since it is a clique already)

The size of expanded matrix is based on the number of cliques. assuming if there is two cliques in the graph, then we have 8 binary variables (Xij, 1 if node i assigned to clique j, 0 otherwise) and the expanded matrix goes to 8 by 8. If there is three cliques, then the matrix goes to 12 by 12, etc.




"Matt J " <mattjacREMOVE(a)THISieee.spam> wrote in message <hnimgq$b8u$1(a)fred.mathworks.com>...
> "Yaquan Xu" <yaquanxu(a)gmail.com> wrote in message <hni6u4$pb2$1(a)fred.mathworks.com>...
> > Thank you so much. It really solved my problem!
>
> But you might want to explain a bit what you need this for.
> It's important to point out that one rarely needs to explicitly expand a matrix using
> kron(A,B), as you're doing. It's usually much more efficient to work with A and B separately, which is why I made this tool:
>
> http://www.mathworks.com/matlabcentral/fileexchange/25969-efficient-object-oriented-kronecker-product-manipulation