pack

Arguments

A

m-by-n real matrix of doubles, the entries of a list of block-diagonal symmetric matrices. n is the number of matrices stored into A.

blocksizes

b-by-1 real matrix of doubles, the sizes of the blocks.

CA

p-by-n real matrix of doubles, a compressed representation of A.

sel

1-by-s real matrix of doubles, the indices of the rows of A which have been selected in CA.

Description

This function takes as input argument a list of block-diagonal matrices stored in the m-by-n matrix A. Only the non-zero entries of the block-diagonal matrices are stored. The integer n is the number of block-diagonal matrices, while the integer m is the number of nonzero entries of one single block-diagonal matrix. The function removes the symmetric entries, and returns in CA the compressed representation of A.

In the CA matrix, the symmetric entries stored in A have been removed. The rows which have been selected in CA are stored in the vector sel, so that, on output, we have CA == A(sel,:).

For example, the matrix

is stored as

            [1; 2; 3; 4; 5; 6; 7; 8; 9; 10]
        

with blocksizes=[2,3,1].

This function is designed to be used when preparing the input arguments of the semidef function.

Examples

In the following example, we compress a single block-diagonal symmetric matrix Z. This is the example presented in "SP: Software for Semidefinite Programming, User's Guide, Beta Version", November 1994, L. Vandenberghe and S. Boyd, 1994, on page 5.

Z = [
1 2 0 0 0 0
2 3 0 0 0 0
0 0 4 5 6 0
0 0 5 7 8 0
0 0 6 8 9 0
0 0 0 0 0 10
];
blocksizes=[2,3,1];
Z1 = Z(1:2,1:2);
Z2 = Z(3:5,3:5);
Z3 = Z(6,6);
A = list2vec(list(Z1,Z2,Z3));
[CA,sel] = pack(A,blocksizes)

In the following example, we compress three block-diagonal symmetric matrices F0, F1 and F2.

// Define 3 symmetric block-diagonal matrices: F0, F1, F2
F0=[2,1,0,0;
    1,2,0,0;
    0,0,3,1;
    0,0,1,3]
F1=[1,2,0,0;
    2,1,0,0;
    0,0,1,3;
    0,0,3,1]
F2=[2,2,0,0;
    2,2,0,0;
    0,0,3,4;
    0,0,4,4]
// Define the size of the two blocks:
// the first block has size 2,
// the second block has size 2.
blocksizes=[2,2];
// Extract the two blocks of the matrices.
F01=F0(1:2,1:2);
F02=F0(3:4,3:4);
F11=F1(1:2,1:2);
F12=F1(3:4,3:4);
F21=F2(1:2,1:2);
F22=F2(3:4,3:4);
// Create 3 column vectors, containing nonzero entries
// in F0, F1, F2.
F0nnz = [F01(:);F02(:)];
F1nnz = [F11(:);F12(:)];
F2nnz = [F21(:);F22(:)];
// Create a 16-by-3 matrix, representing the
// nonzero entries of the 3 matrices F0, F1, F2.
A=[F0nnz,F1nnz,F2nnz]
// Pack the list of matrices A into CA.
[CA,sel] = pack(A,blocksizes)
// Check that CA == A(sel,:)
A(sel,:)

References

L. Vandenberghe and S. Boyd, " Semidefinite Programming," Informations Systems Laboratory, Stanford University, 1994.

Ju. E. Nesterov and M. J. Todd, "Self-Scaled Cones and Interior-Point Methods in Nonlinear Programming," Working Paper, CORE, Catholic University of Louvain, Louvain-la-Neuve, Belgium, April 1994.

SP: Software for Semidefinite Programming, User's Guide, Beta Version, November 1994, L. Vandenberghe and S. Boyd, 1994 http://www.ee.ucla.edu/~vandenbe/sp.html

See also

• unpack — Uncompress a list of block-diagonal symmetric matrices.