Advertisement

Deterministic Discrepancy Minimization via the Multiplicative Weight Update Method

  • Avi Levy
  • Harishchandra Ramadas
  • Thomas Rothvoss
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 10328)

Abstract

A well-known theorem of Spencer shows that any set system with n sets over n elements admits a coloring of discrepancy \(O(\sqrt{n})\). While the original proof was non-constructive, recent progress brought polynomial time algorithms by Bansal, Lovett and Meka, and Rothvoss. All those algorithms are randomized, even though Bansal’s algorithm admitted a complicated derandomization.

We propose an elegant deterministic polynomial time algorithm that is inspired by Lovett-Meka as well as the Multiplicative Weight Update method. The algorithm iteratively updates a fractional coloring while controlling the exponential weights that are assigned to the set constraints.

A conjecture by Meka suggests that Spencer’s bound can be generalized to symmetric matrices. We prove that \(n \times n\) matrices that are block diagonal with block size q admit a coloring of discrepancy \(O(\sqrt{n} \cdot \sqrt{\log (q)})\). Bansal, Dadush and Garg recently gave a randomized algorithm to find a vector x with entries in \(\lbrace {-1,1\rbrace }\) with \(\Vert Ax\Vert _{\infty } \le O(\sqrt{\log n})\) in polynomial time, where A is any matrix whose columns have length at most 1. We show that our method can be used to deterministically obtain such a vector.

Keywords

Symmetric Matrix Polynomial Time Algorithm Symmetric Matrice Semidefinite Program Exponential Weight 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

References

  1. [AHK12]
    Arora, S., Hazan, E., Kale, S.: The multiplicative weights update method: a meta-algorithm and applications. Theor. Comput. 8(6), 121–164 (2012)MathSciNetCrossRefMATHGoogle Scholar
  2. [AS08]
    Alon, N., Spencer, J.H.: The Probabilistic Method. Wiley-Interscience Series in Discrete Mathematics and Optimization, 3rd edn. John Wiley & Sons Inc., Hoboken (2008). With an appendix on the life and work of Paul ErdősCrossRefMATHGoogle Scholar
  3. [Ban98]
    Banaszczyk, W.: Balancing vectors and Gaussian measures of \(n\)-dimensional convex bodies. Random Struct. Algorithms 12(4), 351–360 (1998)MathSciNetCrossRefMATHGoogle Scholar
  4. [Ban10]
    Bansal, N.: Constructive algorithms for discrepancy minimization. In: FOCS, pp. 3–10 (2010)Google Scholar
  5. [BDG16]
    Bansal, N., Dadush, D., Garg, S: An algorithm for komlós conjecture matching banaszczyk’s bound. CoRR, abs/1605.02882 (2016)Google Scholar
  6. [Bec81]
    Beck, J.: Roth’s estimate of the discrepancy of integer sequences is nearly sharp. Combinatorica 1(4), 319–325 (1981)MathSciNetCrossRefMATHGoogle Scholar
  7. [BF81]
    Beck, J., Fiala, T.: “Integer-making" theorems. Discrete Appl. Math. 3(1), 1–8 (1981)MathSciNetCrossRefMATHGoogle Scholar
  8. [Bha97]
    Rajendra, B.: Matrix Analysis. Graduate Texts in Mathematics. Springer, New York (1997)MATHGoogle Scholar
  9. [Boh90]
    Bohus, G.: On the discrepancy of 3 permutations. Random Struct. Algorithms 1(2), 215–220 (1990)MathSciNetCrossRefMATHGoogle Scholar
  10. [BS13]
    Bansal, N., Spencer, J.: Deterministic discrepancy minimization. Algorithmica 67(4), 451–471 (2013)MathSciNetCrossRefMATHGoogle Scholar
  11. [Cha01]
    Chazelle, B.: The Discrepancy Method - Randomness and Complexity. University Press, Cambridge (2001)MATHGoogle Scholar
  12. [Gia97]
    Giannopoulos, A.: On some vector balancing problems. Stud. Math. 122(3), 225–234 (1997)MathSciNetMATHGoogle Scholar
  13. [Glu89]
    Gluskin, E.D.: Extremal properties of orthogonal parallelepipeds and their applications to the geometry of banach spaces. Math. USSR Sb. 64(1), 85 (1989)MathSciNetCrossRefMATHGoogle Scholar
  14. [LM12]
    Lovett, S., Meka, R.: Constructive discrepancy minimization by walking on the edges. In: FOCS, pp. 61–67 (2012)Google Scholar
  15. [Mat99]
    Matoušek, J.: Geometric Discrepancy. Algorithms and Combinatorics. Springer, Berlin (1999). An illustrated guideCrossRefMATHGoogle Scholar
  16. [MSS13]
    Marcus, A., Spielman, D.A., Srivastava, N.: Interlacing families, I.I.: mixed characteristic polynomials and the Kadison-singer problem. arXiv e-prints, June 2013Google Scholar
  17. [Nik13]
    Nikolov, A.: The komlos conjecture holds for vector colorings. arXiv e-prints, Jan 2013Google Scholar
  18. [NNN12]
    Newman, A., Neiman, O., Nikolov, A.: Beck’s three permutations conjecture: a counterexample and some consequences. In: FOCS, pp. 253–262 (2012)Google Scholar
  19. [Spe77]
    Spencer, J.: Balancing games. J. Comb. Theor. Ser. B 23(1), 68–74 (1977)MathSciNetCrossRefMATHGoogle Scholar
  20. [Spe85]
    Spencer, J.: Six standard deviations suffice. Trans. Am. Math. Soc. 289(2), 679–706 (1985)MathSciNetCrossRefMATHGoogle Scholar
  21. [Sri97]
    Srinivasan, A.: Improving the discrepancy bound for sparse matrices: Better approximations for sparse lattice approximation problems. In: SODA 1997, ACM SIGACT, SIAM, Philadelphia, PA, pp. 692–701 (1997)Google Scholar
  22. [SST]
    Spencer, J.H., Srinivasan, A., Tetali, P.: The discrepancy of permutation families. Unpublished manuscriptGoogle Scholar
  23. [Zou12]
    Zouzias, A.: A Matrix hyperbolic cosine algorithm and applications. In: Czumaj, A., Mehlhorn, K., Pitts, A., Wattenhofer, R. (eds.) ICALP 2012. LNCS, vol. 7391, pp. 846–858. Springer, Heidelberg (2012). doi: 10.1007/978-3-642-31594-7_71 CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Avi Levy
    • 1
  • Harishchandra Ramadas
    • 1
  • Thomas Rothvoss
    • 1
  1. 1.University of WashingtonSeattleUSA

Personalised recommendations