Abstract
Consider a matrix with m rows and n pairs of columns. The linear matroid parity problem (LMPP) is to determine a maximum number of pairs of columns that are linearly independent. We show how to solve the linear matroid parity problem as a sequence of matroid intersection problems. The algorithm runs in . Our algorithm is comparable to the best running time for the LMPP, and is far simpler and faster than the algorithm of Orlin and Vande Vate [10], who also solved the LMPP as a sequence of matroid intersection problems. In addition, the algorithm may be viewed naturally as an extension of the blossom algorithm for nonbipartite matchings. © 2008 Springer-Verlag Berlin Heidelberg.
Author supplied keywords
Cite
CITATION STYLE
Orlin, J. B. (2008). A fast, simpler algorithm for the matroid parity problem. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5035 LNCS, pp. 240–258). https://doi.org/10.1007/978-3-540-68891-4_17
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
