Piecewise Convex–Concave Approximation in the Minimax Norm

  • Michael P. Cullinan
Part of the Springer Optimization and Its Applications book series (SOIA, volume 145)


Suppose that \(\mathbf {f} \in \mathbb {R}^{n}\) is a vector of n error-contaminated measurements of n smooth function values measured at distinct, strictly ascending abscissæ. The following projective technique is proposed for obtaining a vector of smooth approximations to these values. Find y minimizing ∥y −f subject to the constraints that the consecutive second-order divided differences of the components of y change sign at most q times. This optimization problem (which is also of general geometrical interest) does not suffer from the disadvantage of the existence of purely local minima and allows a solution to be constructed in only \(O(nq \log n)\) operations. A new algorithm for doing this is developed and its effectiveness is proved. Some results of applying it to undulating and peaky data are presented, showing that it is fast and can give very good results, particularly for large densely packed data, even when the errors are quite large.


  1. 1.
    Cullinan, M.P.: Data smoothing by adjustment of divided differences. Ph.D. dissertation, Department of Applied Mathematics and Theoretical Physics, University of Cambridge (1986)Google Scholar
  2. 2.
    Cullinan, M.P.: Data smoothing using non-negative divided differences and 2 approximation. IMA J. Numer. Anal. 10, 583–608 (1990)MathSciNetCrossRefGoogle Scholar
  3. 3.
    Cullinan, M.P., Powell, M.J.D.: Data smoothing by divided differences. In: Watson, G.A. (ed.) Numerical Analysis Proceedings, Dundee 1981. Lecture Notes in Mathematics, vol. 912, pp. 26–37. Springer, Berlin (1982)Google Scholar
  4. 4.
    Demetriou, I.C.: Signs of divided differences yield least squares data fitting with constrained monotonicity or convexity. J. Comput. Appl. Math. 146(2), 179–211 (2002)MathSciNetCrossRefGoogle Scholar
  5. 5.
    Demetriou, I.C. Powell, M.J.D.: Least squares fitting to univariate data subject to restrictions on the signs of the second differences. In: Iserles, A. Buhmann, M.D. (eds.) Approximation Theory and Optimization. Tributes to M. J. D. Powell, pp. 109–132. Cambridge University Press, Cambridge (1997)Google Scholar
  6. 6.
    Graham, R.L.: An efficient algorithm for determining the convex hull of a finite point set. Inform. Proc. Lett. 1, 132–133 (1972)CrossRefGoogle Scholar
  7. 7.
    Hildebrand, F.B.: Introduction to Numerical Analysis. McGraw Hill, New York (1956)zbMATHGoogle Scholar
  8. 8.
    Kruskal, J.B.: Non-metric multidimensional scaling: a numerical method. Psychometrika 29, 115–129 (1964)MathSciNetCrossRefGoogle Scholar
  9. 9.
    Miles, R.E.: The complete amalgamation into blocks, by weighted means, of a finite set of real numbers. Biometrika 46, 317–327 (1959)MathSciNetCrossRefGoogle Scholar
  10. 10.
    Petit, J.R., et al.: Vostok Ice Core Data for 420,000 Years, IGBP PAGES/World Data Center for Paleoclimatology Data Contribution Series #2001-076. NOAA/NGDC Paleoclimatology Program, Boulder (2001)Google Scholar
  11. 11.
    Ubhaya, V.A.: An O(n) algorithm for discrete n-point convex approximation with applications to continuous case. J. Math. Anal. Appl. 72(1), 338–354 (1979)MathSciNetCrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Michael P. Cullinan
    • 1
  1. 1.Maryvale InstituteBirminghamUK

Personalised recommendations