Advertisement

Irregularities of Distribution, Derandomization, and Complexity Theory

  • Bernard Chazelle
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1974)

Abstract

In 1935, van der Corput asked the following question: Given an infinite sequence of reals in [0, 1], define
$$ D(n) = \mathop {\sup }\limits_{o \leqslant x \leqslant 1} \left| {|S_n \cap [0,x]| - nx} \right|,$$
where Sn consists of the first n elements in the sequence. Is it possible for D(n) to stay in O(1)? Many years later, Schmidt proved that D(n) can never be in o(log n). In other words, there are limitations on how well the discrete distribution, x → ∣S n ∩ [0, x] |, can simulate the continuous one, x → nx. The study of this intriguing phenomenon and its numerous variants related to the irregularities of distributions has given rise to discrepancy theory. The relevance of the subject to complexity theory is most evident in the study of probabilistic algorithms. Suppose that we feed a probabilistic algorithm not with a perfectly random sequence of bits (as is usually required) but one that is only pseudorandom or even deterministic. Should performance necessarily suffer? In particular, suppose that one could trade an exponential-size probability space for one of polynomial size without letting the algorithm realize the change. This form of derandomization can be expressed by saying that a very large distribution can be simulated by a small one for the purpose of the algorithm. Put differently, there exists a measure with respect to which the two distributions have low discrepancy. The study of discrepancy theory predates complexity theory and a wealth of mathematical techniques can be brought to bear to prove nontrivial derandomization results. The pipeline of ideas that flows from discrepancy theory to complexity theory constitutes the discrepancy method. We give a few examples in this survey. A more thorough treatment is given in our book [15]. We also briefly discuss the relevance of the discrepancy method to complexity lower bounds.

Keywords

Complexity Theory Discrepancy Theory Circuit Complexity Probabilistic Algorithm Monotone Circuit 
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.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Agarwal, P.K. Partitioning arrangements of lines II: Applications, Disc. Comput. Geom. 5 (1990), 533–573.zbMATHCrossRefGoogle Scholar
  2. 2.
    Agarwal, P.K. Geometric partitioning and its applications, in Computational Geometry: Papers from the DIMACS Special Year, eds., Goodman, J.E., Pollack, R., Steiger, W., Amer. Math. Soc., 1991.Google Scholar
  3. 3.
    Agarwal, P.K., Erickson, J. Geometric range searching and its relatives, in Advances in Discrete and Computational Geometry, eds. Chazelle, B., Goodman, J.E., Pollack, R., Contemporary Mathematics 223, Amer. Math. Soc., 1999, pp. 1–56.Google Scholar
  4. 4.
    Alon, N., Spencer, J.H. The Probabilistic Method, Wiley-Interscience, 1992.Google Scholar
  5. 5.
    Beck, J. Irregularities of distribution, I, Acta Math. 159 (1987), 1–49.zbMATHCrossRefMathSciNetGoogle Scholar
  6. 6.
    Beck, J., Chen, W.W.L. Irregularities of Distribution, Cambridge Tracts in Mathematics, 89, Cambridge University Press, 1987.Google Scholar
  7. 7.
    Beck, J., Fiala, T. “Integer-making” theorems, Discrete Applied Mathematics 3(1981), 1–8.zbMATHCrossRefMathSciNetGoogle Scholar
  8. 8.
    Beck, J., Sós, V.T. Discrepancy theory, in Handbook of Combinatorics, Chap. 26, eds., Graham, R.L., Grötschel, M., Lovász, L., North-Holland, 1995, pp. 1405–1446.Google Scholar
  9. 9.
    Chazelle, B. Lower bounds on the complexity of polytope range searching, J. Amer. Math. Soc. 2 (1989), 637–666.zbMATHCrossRefMathSciNetGoogle Scholar
  10. 10.
    Chazelle, B. Lower bounds for orthogonal range searching: II. The arithmetic model, J. ACM37 (1990), 439–463.zbMATHCrossRefMathSciNetGoogle Scholar
  11. 11.
    Chazelle, B. Cutting hyperplanes for divide-and-conquer, Disc. Comput. Geom. 9 (1993), 145–158.zbMATHCrossRefMathSciNetGoogle Scholar
  12. 12.
    Chazelle, B. An optimal convex hull algorithm in any fixed dimension, Disc. Comput. Geom. 10 (1993), 377–409.zbMATHCrossRefMathSciNetGoogle Scholar
  13. 13.
    Chazelle, B. Lower bounds for off-line range searching, Disc. Comput. Geom. 17 (1997), 53–65.zbMATHCrossRefMathSciNetGoogle Scholar
  14. 14.
    Chazelle, B. A spectral approach to lower bounds with applications to geometric searching, SIAM J. Comput. 27 (1998), 545–556.zbMATHCrossRefMathSciNetGoogle Scholar
  15. 15.
    Chazelle, B. The Discrepancy Method: Randomness and Complexity, Cambridge University Press, 2000.Google Scholar
  16. 16.
    Chazelle, B., Friedman, J. A deterministic view of random sampling and its use in geometry, Combinatorica 10 (1990), 229–249.zbMATHCrossRefMathSciNetGoogle Scholar
  17. 17.
    Chazelle, B., Lvov, A. A trace bound for the hereditary discrepancy, Proc. 16th Annual ACM Symp. Comput. Geom. (2000), 64–69. To appear in Disc. Comput. Geom.Google Scholar
  18. 18.
    Chazelle, B., Matoušek, J. On linear-time deterministic algorithms for optimization problems in fixed dimension, J. Algorithms 21 (1996), 579–597.zbMATHCrossRefMathSciNetGoogle Scholar
  19. 19.
    Clarkson, K.L. Linear programming in O(n * 3d/2 ) time, Inform. Process. Lett. 22 (1986), 21–24.CrossRefMathSciNetGoogle Scholar
  20. 20.
    Clarkson, K.L. New applications of random sampling in computational geometry, Disc. Comput. Geom. 2 (1987), 195–222.zbMATHCrossRefMathSciNetGoogle Scholar
  21. 21.
    Clarkson, K.L. Las Vegas algorithms for linear and integer programming when the dimension is small, J. ACM 42 (1995), 488-499.Google Scholar
  22. 22.
    Davenport, H. Note on irregularities of distribution, Mathematika 3 (1956), 131–135.zbMATHMathSciNetGoogle Scholar
  23. 23.
    Dyer, M.E. On a multidimensional search technique and its application to the Euclidean one-centre problem, SIAM J. Comput. 15 (1986), 725–738.zbMATHCrossRefMathSciNetGoogle Scholar
  24. 24.
    Dyer, M.E., Frieze, A.M. A randomized algorithm for fixed-dimensional linear programming,tiMathematical Programming 44 (1989), 203–212.Google Scholar
  25. 25.
    Hammersley, J.M. Monte Carlo methods for solving multivariable problems, Ann. New York Acad. Sci. 86 (1960), 844–874.zbMATHCrossRefMathSciNetGoogle Scholar
  26. 26.
    Haussler, D., Welzl, E. ∈-nets and simplex range queries, Disc. Comput. Geom. 2 (1987), 127–151.zbMATHCrossRefMathSciNetGoogle Scholar
  27. 27.
    Matoušek, J. Construction of ∈-nets, Disc. Comput. Geom. 5 (1990), 427–448.CrossRefzbMATHGoogle Scholar
  28. 28.
    Matoušek, J. Geometric range searching, ACM Comput. Surv. 26 (1994), 421–461.Google Scholar
  29. 29.
    Matoušek, J. Approximations and optimal geometric divide-and-conquer, J. Comput. Syst. Sci. 50 (1995), 203–208.CrossRefGoogle Scholar
  30. 30.
    Matoušek, J. Derandomization in computational geometry, J. Algorithms 20 (1996), 545–580.CrossRefzbMATHMathSciNetGoogle Scholar
  31. 31.
    Matoušek, J. Geometric Discrepancy: An Illustrated Guide, Algorithms and Combinatorics, 18, Springer, 1999.Google Scholar
  32. 32.
    Megiddo, N. Linear-time algorithms for linear programming in R3 and related problems, SIAM J. Comput. 12 (1983), 759–776.zbMATHCrossRefMathSciNetGoogle Scholar
  33. 33.
    Megiddo, N. Linear programming in linear time when the dimension is fixed, J. ACM31 (1984), 114–127.zbMATHCrossRefMathSciNetGoogle Scholar
  34. 34.
    Montgomery, H.L. On irregularities of distribution, in Congress of Number Theory (Zarautz, 1984), Universidad del País Vasco, Bilbao, 1989, pp. 11–27.Google Scholar
  35. 35.
    Montgomery, H.L. Ten Lectures on the Interface Between Analytic Number Theory and Harmonic Analysis, CBMS Regional Conference Series in Mathematics, No. 84, Amer. Math. Soc., Providence, 1994.Google Scholar
  36. 36.
    Motwani, R., Raghavan, P. Randomized Algorithms, Cambridge University Press, 1995.Google Scholar
  37. 37.
    Niederreiter, H. Random Number Generation and Quasi-Monte Carlo Methods, CBMS-NSF, SIAM, Philadelphia, PA, 1992.zbMATHGoogle Scholar
  38. 38.
    Pach, J., Agarwal, P.K. Combinatorial Geometry, Wiley-Interscience Series in Discrete Mathematics and Optimization, John Wiley & Sons, Inc., 1995.Google Scholar
  39. 39.
    Roth, K.F. On irregularities of distribution, Mathematika 1 (1954), 73–79.zbMATHMathSciNetCrossRefGoogle Scholar
  40. 40.
    Roth, K.F. Remarkc oncerning integer sequences, Acta Arithmetica 9 (1964), 257–260.zbMATHMathSciNetGoogle Scholar
  41. 40.
    Roth, K.F. Remarkc oncerning integer sequences, Acta Arithmetica 9 (1964), 257–260.zbMATHMathSciNetGoogle Scholar
  42. 41.
    Schmidt, W.M. Irregularities of distribution, VII, Acta Arithmetica 21 (1972), 45–50.zbMATHMathSciNetGoogle Scholar
  43. 42.
    Seidel, R. Small-dimensional linear programming and convex hulls made easy, Disc. Comput. Geom. 6 (1991), 423–434.zbMATHCrossRefMathSciNetGoogle Scholar
  44. 43.
    Sharir, M., Welzl, E. A combinatorial bound for linear programming and related problems, Proc. 9th Annual Symp. Theoret. Aspects Comput. Sci., LNCS, 577, Springer-Verlag, 1992, pp. 569–579.Google Scholar
  45. 44.
    Spencer, J. Six standard deviations suffice, Trans. Amer. Math. Soc. 289 (1985), 679–706.zbMATHCrossRefMathSciNetGoogle Scholar
  46. 45.
    Spencer, J. Ten Lectures on the Probabilistic Method, CBMS-NSF, SIAM, 1987.Google Scholar
  47. 46.
    van der Corput, J.G. Verteilungsfunktionen I. Proc. Nederl. Akad. Wetensch. 38 (1935), 813–821.Google Scholar
  48. 47.
    van der Corput, J.G. Verteilungsfunktionen II. Proc. Nederl. Akad. Wetensch. 38 (1935), 1058–1066.Google Scholar
  49. 48.
    Vapnik, V.N., Chervonenkis, A.Ya. On the uniform convergence of relative frequencies of events to their probabilities, Theory of Probability and its Applications 16 (1971), 264–280.CrossRefzbMATHGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2000

Authors and Affiliations

  • Bernard Chazelle
    • 1
  1. 1.Department of Computer SciencePrinceton University, and NEC Research InstitutePrinceton

Personalised recommendations