Semidefinite programming (SDP) may be viewed as an extension of linear programming (LP), and most interior point methods (IPM s) for LP can be extended to solve SDP problems. However, it is far more difficult to exploit data structures (especially sparsity) in the SDP case. In this paper we will look at the data structure where the SDP data matrices lie in a low dimensional matrix algebra. This data structure occurs in several applications, including the lower bounding of the stability number in certain graphs and the crossing number in complete bipartite graphs. We will show that one can reduce the linear algebra involved in an iteration of an IPM to involve matrices of the size of the dimension of the matrix algebra only. In other words, the original sizes of the data matrices do not appear in the computational complexity bound. In particular, we will work out the details for the dual scaling algorithm, since a dual method is most suitable for the types of applications we have in mind.
Download Info
To download:
If you experience problems downloading a file, check if you have the
proper application to
view it first. Information about this may be contained
in the File-Format links below. In case of further problems read
the IDEAS help
page. Note that these files are not on the IDEAS
site. Please be patient as the files may be large.
Publisher Info
Paper provided by Tilburg University, Center for Economic Research in its series Discussion Paper with number
17.
References listed on IDEAS Please report citation or reference errors to , or , if you are the registered author of the cited work, log in to your RePEc Author Service profile, click on "citations" and make appropriate adjustments.: