A characterization theorem and an algorithm for a convex hull problem

Abstract

Given \(S= \{v_1, \dots , v_n\} \subset {\mathbb {R}}^m\) and \(p \in {\mathbb {R}}^m\), testing if \(p \in conv(S)\), the convex hull of \(S\), is a fundamental problem in computational geometry and linear programming. First, we prove a Euclidean distance duality, distinct from classical separation theorems such as Farkas Lemma: \(p\) lies in \(conv(S)\) if and only if for each \(p' \in conv(S)\) there exists a pivot, \(v_j \in S\) satisfying \(d(p',v_j) \ge d(p,v_j)\). Equivalently, \(p \not \in conv(S)\) if and only if there exists a witness, \(p' \in conv(S)\) whose Voronoi cell relative to \(p\) contains \(S\). A witness separates \(p\) from \(conv(S)\) and approximate \(d(p, conv(S))\) to within a factor of two. Next, we describe the Triangle Algorithm: given \(\epsilon \in (0,1)\), an iterate, \(p' \in conv(S)\), and \(v \in S\), if \(d(p, p') < \epsilon d(p,v)\), it stops. Otherwise, if there exists a pivot \(v_j\), it replace \(v\) with \(v_j\) and \(p'\) with the projection of \(p\) onto the line \(p'v_j\). Repeating this process, the algorithm terminates in \(O(mn \min \{ \epsilon ^{-2}, c^{-1}\ln \epsilon ^{-1} \})\) arithmetic operations, where \(c\) is the visibility factor, a constant satisfying \(c \ge \epsilon ^2\) and \(\sin (\angle pp'v_j) \le 1/\sqrt{1+c}\), over all iterates \(p'\). In particular, the geometry of the input data may result in efficient complexities such as \(O(mn \root t \of {\epsilon ^{-2}} \ln \epsilon ^{-1})\), \(t\) a natural number, or even \(O(mn \ln \epsilon ^{-1})\). Additionally, (i) we prove a strict distance duality and a related minimax theorem, resulting in more effective pivots; (ii) describe \(O(mn \ln \epsilon ^{-1})\)-time algorithms that may compute a witness or a good approximate solution; (iii) prove generalized distance duality and describe a corresponding generalized Triangle Algorithm; (iv) prove a sensitivity theorem to analyze the complexity of solving LP feasibility via the Triangle Algorithm. The Triangle Algorithm is practical and competitive with the simplex method, sparse greedy approximation and first-order methods. Finally, we discuss future work on applications and generalizations.

This is a preview of subscription content, log in to check access.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13
Fig. 14
Fig. 15
Fig. 16
Fig. 17

References

  1. Asano, T., Matoušek, J., & Tokuyama, T. (2007). Zone diagrams: Existence, uniqueness, and algorithmic challenge. SIAM Journal on Computing, 37, 1182–1198.

    Article  Google Scholar 

  2. Aurenhammer, F. (1991). Voronoi diagrams—a survey of fundamental geometric data structure. ACM Computing Surveys, 23, 345–405.

    Article  Google Scholar 

  3. Berkowitz, R., Kalantari, B., Memendendez, D., & Kalantari, I. (2012). On properties of forbidden zone of polyhedrons and polytopes. In Proceedings of the ninth annual international symposium on voronoi diagrams in science and engineering, pp. 56–65.

  4. Burges, C. (1998). A tutorial on support vector machines for pattern recognition. Data Mining and Knowledge Discovery, 2, 121–167.

    Article  Google Scholar 

  5. Chan, T. M. (1996). Output-sensitive results on convex hulls, extreme points, and related problems. Discrete & Computational Geometry, 16, 369–387.

    Article  Google Scholar 

  6. Chazelle, B. (1993). An optimal convex hull algorithm in any fixed dimension. Discrete & Computational Geometry, 10, 377–409.

    Article  Google Scholar 

  7. Chvátal, V. (1983). Linear programming. New York: W.H. Freeman and Company.

    Google Scholar 

  8. Clarkson, K. L. (1988). Las Vegas algorithms for linear and integer programming when the dimension is small. In: Proceedings of 29th annual IEEE symposium on foundations of computer science, pp. 452–456.

  9. Clarkson, K. L. (1994). More output-sensitive geometric algorithm. In Proceedings of the 35-th annual IEEE symposium on foundations of computer science, pp. 695–702.

  10. Clarkson, K. L. (2008). Coresets, sparse greedy approximation, and the Frank-Wolfe algorithm. In SODA ’08: Proceedings of the nineteenth annual ACM-SIAM symposium on discrete algorithms (pp. 922–931). Philadelphia: Society for Industrial and Applied Mathematics.

  11. Dantzig, G. B. (1963). Linear programming and extensions. Princeton, NJ: Princeton University Press.

    Google Scholar 

  12. Dantzig, G. B. (1992). An \(\epsilon \)-precise feasible solution to a linear program with a convexity constraint in \(1/\epsilon ^2\) iterations independent of problem size. Technical Report, Stanford University.

  13. de Biasi, S. C., Kalantari, B., & Kalantari, I. (2010). Maximal zone diagrams and their computation. In Proceedings of the seventh annual international symposium on voronoi diagrams in science and engineering, pp. 171–180.

  14. de Biasi, S. C., Kalantari, B., & Kalantari, I. (2011). Mollified zone diagrams and their computation. In Transactions on computational science XIV: Lecture notes in computer science (Vol. 6970, pp. 31–59).

  15. Durier, R., & Michelot, C. (1986). Sets of efficient points in a normed space. Journal of Mathematical Analysis and Applications, 117, 506–528.

    Article  Google Scholar 

  16. Dyer, M. E., & Frieze, A. M. (1989). A randomized algorithm for fixed-dimensional linear programming. Mathematical Programming, 44, 203–212.

    Article  Google Scholar 

  17. Epelman, M., & Freund, R. M. (1997). Condition number complexity of an elementary algorithm for resolving a conic linear system, Technical Report. http://dspace.mit.edu/bitstream/handle/1721.1/5254/or-319-97-37692575?sequence=1.

  18. Frank, M., & Wolfe, P. (1956). An algorithm for quadratic programming. Naval Research Logistics Quarterly, 3, 95–110.

    Article  Google Scholar 

  19. Gärtner, B., & Jaggi, M. (2009). Coresets for polytope distance. In Symposium on computational geometry, pp. 33–42.

  20. Gibson, T., & Kalantari, B. (2013). Experiments with the triangle algorithm for linear systems, 2-page Extended Abstract, 23nd annual fall workshop on computational geometry, City College of New York.

  21. Gilbert, E. G. (1966). An iterative procedure for computing the minimum of a quadratic form on a convex set. SIAM Journal on Control, 4, 61–80.

    Article  Google Scholar 

  22. Goodman, J. E., & O’Rourke, J. (Eds.). (2004). Handbook of discrete and computational geometry, 2nd Edn (Discrete Mathematics and Its Applications). Boca Raton: Chapman & Hall.

    Google Scholar 

  23. Har-Peled, S., Roth, D., & Zimak, D. (2007). Maximum margin coresets for active and noise tolerant learning. IJCAI.

  24. Jin, Y., & Kalantari, B. (2006). A procedure of Chvátal for testing feasibility in linear programming and matrix scaling. Linear Algebra and its Applications, 416, 795–798.

    Article  Google Scholar 

  25. Kalai, G. (1992). A subexponential randomized simplex algorithm. In Proceedings of the 24-th annual ACM symposium on theory of computing, pp. 475–482.

  26. Kalantari, B. (2014). An algorithmic separating hyperplane theorem and its applications.

  27. Kalantari, B. (1990). Canonical problems for quadratic programming and projective methods for their solution. Contemporary Mathematics, 114, 243–263.

    Article  Google Scholar 

  28. Kalantari, B. (2003). Semidefinite programming and matrix scaling over the semidefinite cone. Linear Algebra and its Applications, 375, 221–243.

    Article  Google Scholar 

  29. Kalantari, B. (2012a). A characterization theorem and an algorithm for a convex hull problem. arXiv:1204.1873v2.

  30. Kalantari, B. (2012b). Finding a lost treasure in convex hull of points from known distances. In The Proceedings of the 24th Canadian conference on computational geometry, pp. 271–276.

  31. Kalantari, B. (2012c). Solving linear system of equations via a convex hull algorithm. arXiv:1210.7858v1.

  32. Kalantari, B., & Emamy-K, M. R. (1997). On linear programming and matrix scaling over the algebraic numbers. Linear Algebra and its Applications, 262, 283–306.

    Article  Google Scholar 

  33. Kalantari, B., & Saks, M. (2013). On the triangle algorithm for the convex hull membership, 2-page Extended Abstract, 23nd annual fall workshop on computational geometry, City College of New York.

  34. Karmarkar, N. (1984). A new polynomial time algorithm for linear programming. Combinatorica, 4, 373–395.

    Article  Google Scholar 

  35. Kelner, J. A., & Spielman, D. A. (2006). A randomized polynomial-time simplex algorithm for linear programming, In Proceedings of the 38th annual ACM symposium on theory of computing, pp 51–60.

  36. Khachiyan, L. G. (1979). A polynomial algorithm in linear programming. Doklady Akademia Nauk SSSR, 244, 1093–1096 (translated in Soviet Mathematics Doklady 20, 191–194, 1979).

  37. Khachiyan, L., & Kalantari, B. (1992). Diagonal matrix scaling and linear programming. SIAM Journal on Optimization, 4, 668–672.

    Article  Google Scholar 

  38. Klee, V., & Minty, G. J. (1972). How good is the simplex algorithm? In O. Shisha (Ed.), Inequalities III (pp. 159–175). London: Academic Press.

  39. Kuhn, H. W. (2011). On a pair of dual nonlinear programs. In J. Abadie (Ed.), Nonlinear programming (pp. 37–54). New York: Wiley, 1967, see also, Locational problems and mathematical programming, in Mathematical Optimiation in Economics C.I.M.E. Summer Schools, Vol. 38, pp. 57–76.

  40. Lan, G., Lu, Z., & Monteiro, R. D. C. (2011). Primal-dual first-order methods with \(O(1/\epsilon )\) iteration-complexity for cone programming. Mathematical Programming Series A, 126, 1–29.

    Article  Google Scholar 

  41. Li, M., & Kalantari, B. (2013). Experimental study of the convex hull decision problem via a new geometric algorithm, 2-page Extended Abstract, 23nd annual fall workshop on computational geometry, City College of New York.

  42. Matoušek, J. (1993). Linear optimization queries. Journal of Algorithms, 14, 432–448.

    Article  Google Scholar 

  43. Matoušek, J., Sharir, M., & Welzl, E. (1992). A subexponential bound for linear programming. In Proceedings of the 8th annual ACM symposium on computational geometry, pp. 1–8.

  44. Megiddo, N. (1984). Linear programming in linear time when the dimension is fixed. Journal of the ACM, 31, 114–127.

    Article  Google Scholar 

  45. Motawani, R., & Raghavan, R. (1995). Randomized algorithms. Cambridge: Cambridge University Press.

    Google Scholar 

  46. Nesterov, Y. E. (2013). Private communications.

  47. Nesterov, Y. E. (2005). Smooth minimization of nonsmooth functions. Mathematical Programming, 103, 127–152.

    Article  Google Scholar 

  48. Preparata, F. P., & Shamos, M. I. (1985). Computational geometry: An introduction. New York: Springer.

    Google Scholar 

  49. Rockafellar, R. T. (1997). Convex analysis. Princeton, New Jersey: Princeton University Press. [1970].

    Google Scholar 

  50. Schrijver, A. (1986). Theory of linear and integer programming. New York, Chichester, Brisbane, Toronto and Singapore: Wiley.

    Google Scholar 

  51. Seidel, R. (1991). Small-dimensional linear programming and convex hulls made easy. Discrete Computational Geometry, 6, 423–434.

    Article  Google Scholar 

  52. Shamir, R. (1987). The efficiency of the simplex method: A survey. Management Science, 33, 301–334.

    Article  Google Scholar 

  53. Sharir, M., & Welzl, E. (1992). A combinatorial bound for linear programming and related problems. In Proceedings of the 9th symposium on theoretical aspects of computer science, pp. 569–579.

  54. Todd, M. J. (2002). The many facets of linear programming. Mathematical Programming, 91, 417–436.

    Article  Google Scholar 

  55. Zhang, T. (2003). Sequential greedy approximation for certain convex optimization problems. IEEE Transactions on Information Theory, 49, 682–691.

    Article  Google Scholar 

  56. Zimak, D. A. (2006). Algorithms and Analysis for Multi-Category Classification, Ph.D. thesis, Department of Computer Science, University of Illinois at Urbana-Champaign.

Download references

Acknowledgments

I thank Iraj Kalantari for a discussion on the use of Voronoi diagram argument to prove Theorem 1, resulting in a simpler geometric proof than the one given in an earlier version of this article. I thank Tong Zhang for a discussion regarding the Greedy Algorithm for general convex minimization over a simplex. I also thank Günter Rote for bringing to my attention the work of Kuhn (2011) and Durier and Michelot (1986) (Remark 4). I thank a referee for bringing to my attention Dantzig (1992) technical report on analysis of von Neumann algorithm. Also, when the article was being refereed Mike Grigoriadis brought to my attention Dantzig’s analysis in Epelman and Freund (1997). I thank a second anonymous referee for his comments.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Bahman Kalantari.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Kalantari, B. A characterization theorem and an algorithm for a convex hull problem. Ann Oper Res 226, 301–349 (2015). https://doi.org/10.1007/s10479-014-1707-2

Download citation

Keywords

  • Convex hull
  • Linear programming
  • Duality
  • Approximation algorithms
  • Gilbert’s algorithm
  • Frank–Wolfe algorithm
  • Minimax