On the linear classification of even and odd permutation matrices and the complexity of computing the permanent
- 22 Downloads
The problem of linear classification of the parity of permutation matrices is studied. This problem is related to the analysis of complexity of a class of algorithms designed for computing the permanent of a matrix that generalizes the Kasteleyn algorithm. Exponential lower bounds on the magnitude of the coefficients of the functional that classifies the even and odd permutation matrices in the case of the field of real numbers and similar linear lower bounds on the rank of the classifying map for the case of the field of characteristic 2 are obtained.
Keywordspermutation matrix parity permanent linear classification theta function independence number
Unable to display preview. Download preview PDF.
- 1.E. Abbe, N. Alon, and A. S. Bandeira, “Linear Boolean classification, coding and the critical problem”, arXiv:1401.6528v2. 2014.Google Scholar
- 6.H. J. Ryser, Combinatorial Mathematics, Carus Math. Monographs, No. 14 (Math. Assoc. of America, Washington, 1963).Google Scholar
- 8.P. W. Kasteleyn, “Graph theory and crystal physics,” in Graph Theory and Theoretical Physics, Ed. by F. Harary (Academic, London, 1967), pp. 43–110.Google Scholar
- 9.Yu. G. Smetanin and L. G. Khachiyan, “Application of pseudopolynomial algorithms to some combinatorial constrained optimization problems,” Izv. Akad. Nauk. SSSR, Tekh. Kibern., No. 6, 139–144 (1986).Google Scholar
- 10.L. Lovasz, “Semidefinite programs and combinatorial optimization,” in Recent Advances in Algorithms and Combinatorics, Ed. by B. A. Reed and C. L. Sales (Springer, Berlin, 2003), pp. 137–194.Google Scholar