Skip to main content

Computing the Homology of Semialgebraic Sets. I: Lax Formulas

Abstract

We describe and analyze an algorithm for computing the homology (Betti numbers and torsion coefficients) of closed semialgebraic sets given by Boolean formulas without negations over lax polynomial inequalities. The algorithm works in weak exponential time. This means that outside a subset of data having exponentially small measure, the cost of the algorithm is single exponential in the size of the data. All previous algorithms solving this problem have doubly exponential complexity. Our algorithm thus represents an exponential acceleration over state-of-the-art algorithms for all input data outside a set that vanishes exponentially fast.

This is a preview of subscription content, access via your institution.

Fig. 1

References

  1. 1.

    D. Amelunxen and M. Lotz. Average-case complexity without the black swans. J. Complexity, 41:82–101, 2017.

    MathSciNet  MATH  Article  Google Scholar 

  2. 2.

    S. Basu. On bounding the Betti numbers and computing the Euler characteristic of semi-algebraic sets. In Proceedings of the Twenty-Eighth Annual ACM Symposium on Theory of Computing, pages 408–417. ACM, 1996.

  3. 3.

    S. Basu. Computing the first few Betti numbers of semi-algebraic sets in single exponential time. Journal of Symbolic Computation, 41(10):1125–1154, 2006.

    MathSciNet  MATH  Article  Google Scholar 

  4. 4.

    S. Basu, R. Pollack, and M.-F. Roy. On the combinatorial and algebraic complexity of quantifier elimination. Journal of the Assoc. Comput. Mach., 43:1002–1045, 1996.

    MathSciNet  MATH  Article  Google Scholar 

  5. 5.

    S. Basu, R. Pollack, and M.-F. Roy. Computing roadmaps of semi-algebraic sets on a variety. Journal of the Amer. Math. Soc., 33:55–82, 1999.

    MathSciNet  MATH  Google Scholar 

  6. 6.

    C. Beltrán and L.M. Pardo. Smale’s 17th problem: average polynomial time to compute affine and projective solutions. J. Amer. Math. Soc., 22(2):363–385, 2009.

    MathSciNet  MATH  Article  Google Scholar 

  7. 7.

    R. Benedetti and J.-J. Risler. Real algebraic and semi-algebraic sets. Actualités Mathématiques. [Current Mathematical Topics]. Hermann, Paris, 1990.

  8. 8.

    L. Blum, M. Shub, and S. Smale. On a theory of computation and complexity over the real numbers: NP-completeness, recursive functions and universal machines. Bull. Amer. Math. Soc. (N.S.), 21(1):1–46, 1989.

    MathSciNet  MATH  Article  Google Scholar 

  9. 9.

    J. Bochnak, M. Coste, and M.-F. Roy. Real algebraic geometry, volume 36 of Ergebnisse der Mathematik und ihrer Grenzgebiete (3). Springer-Verlag, Berlin, 1998. Translated from the 1987 French original, Revised by the authors.

  10. 10.

    H. Bosse, M. Grötschel, and M. Henk. Polynomial inequalities representing polyhedra. Math. Program., 103(1, Ser. A):35–44, 2005.

    MathSciNet  MATH  Article  Google Scholar 

  11. 11.

    M. Brown. Locally flat imbeddings of topological manifolds. Annals of Mathematics, 75:331–341, 1962.

    MathSciNet  MATH  Article  Google Scholar 

  12. 12.

    P. Bürgisser and F. Cucker. Exotic quantifiers, complexity classes, and complete problems. Found. Comput. Math., 9(2):135–170, 2009.

    MathSciNet  MATH  Article  Google Scholar 

  13. 13.

    P. Bürgisser and F. Cucker. On a problem posed by Steve Smale. Annals of Mathematics, 174:1785–1836, 2011.

    MathSciNet  MATH  Article  Google Scholar 

  14. 14.

    P. Bürgisser and F. Cucker. Condition, volume 349 of Grundlehren der mathematischen Wissenschaften. Springer-Verlag, Berlin, 2013.

  15. 15.

    P. Bürgisser, F. Cucker, and P. Lairez. Computing the homology of basic semialgebraic sets in weak exponential time. J. ACM, 66(1):5:1–5:30, December 2018.

    MathSciNet  MATH  Article  Google Scholar 

  16. 16.

    P. Bürgisser, F. Cucker, and M. Lotz. The probability that a slightly perturbed numerical analysis problem is difficult. Mathematics of Computation, 77:1559–1583, 2008.

    MathSciNet  MATH  Article  Google Scholar 

  17. 17.

    J. Canny. The complexity of robot motion planning, volume 1987 of ACM Doctoral Dissertation Awards. MIT Press, Cambridge, MA, 1988.

  18. 18.

    J. Canny. Computing roadmaps of general semi-algebraic sets. Comput. J., 36(5):504–514, 1993.

    MathSciNet  MATH  Article  Google Scholar 

  19. 19.

    J. Canny, D.Yu. Grigorev, and N.N. Vorobjov. Finding connected components of a semialgebraic set in subexponential time. Appl. Algebra Engrg. Comm. Comput., 2(4):217–238, 1992.

    MathSciNet  Article  Google Scholar 

  20. 20.

    F. Chazal, D. Cohen-Steiner, and A. Lieutier. A sampling theory for compact sets in Euclidean space. In Computational geometry (SCG’06), pages 319–326. ACM, New York, 2006.

  21. 21.

    F. Chazal and A. Lieutier. The “\(\lambda \)-medial axis”. Graphical Models, 67(4):304 – 331, 2005.

    MATH  Article  Google Scholar 

  22. 22.

    F. Chazal and A. Lieutier. Weak feature size and persistant homology: computing homology of solids in \(\mathbb{R}^n\) from noisy data samples. In Computational geometry (SCG’05), pages 255–262. ACM, New York, 2005.

  23. 23.

    G.E. Collins. Quantifier elimination for real closed fields by cylindrical algebraic decomposition, pages 134–183. Lecture Notes in Comput. Sci., Vol. 33. Springer, Berlin, 1975.

  24. 24.

    R. Connelly. A New Proof of Brown’s Collaring Theorem. Proc. Am. Math. Soc., 27:180, 1971.

    MathSciNet  MATH  Google Scholar 

  25. 25.

    F. Cucker. Approximate zeros and condition numbers. J. Complexity, 15:214–226, 1999.

    MathSciNet  MATH  Article  Google Scholar 

  26. 26.

    F. Cucker, T. Krick, G. Malajovich, and M. Wschebor. A numerical algorithm for zero counting. I: Complexity and accuracy. J. Complexity, 24:582–605, 2008.

    MathSciNet  MATH  Article  Google Scholar 

  27. 27.

    F. Cucker, T. Krick, G. Malajovich, and M. Wschebor. A numerical algorithm for zero counting. II: Distance to ill-posedness and smoothed analysis . J. Fixed Point Theory Appl., 6:285–294, 2009.

    MathSciNet  MATH  Article  Google Scholar 

  28. 28.

    F. Cucker, T. Krick, G. Malajovich, and M. Wschebor. A numerical algorithm for zero counting. III: Randomization and condition. Adv. Applied Math., 48:215–248, 2012.

    MathSciNet  MATH  Article  Google Scholar 

  29. 29.

    F. Cucker, T. Krick, and M. Shub. Computing the homology of real projective sets. Found. Comput. Math., 18:929–970, 2018.

    MathSciNet  MATH  Article  Google Scholar 

  30. 30.

    F. Cucker and S. Smale. Complexity estimates depending on condition and round-off error. Journal of the Assoc. Comput. Mach., 46:113–184, 1999.

    MathSciNet  MATH  Article  Google Scholar 

  31. 31.

    J. Demmel. The probability that a numerical analysis problem is difficult. Math. Comp., 50:449–480, 1988.

    MathSciNet  MATH  Article  Google Scholar 

  32. 32.

    A.H. Durfee. Neighborhoods of algebraic sets. Trans. Amer. Math. Soc., 276(2):517–530, 1983.

    MathSciNet  MATH  Article  Google Scholar 

  33. 33.

    H. Edelsbrunner and J.L. Harer. Computational topology: An introduction. American Mathematical Society, Providence, RI, 2010.

    MATH  Google Scholar 

  34. 34.

    H. Federer. Curvature measures. Trans. Amer. Math. Soc., 93:418–491, 1959.

    MathSciNet  MATH  Article  Google Scholar 

  35. 35.

    C.G. Gibson, K. Wirthmüller, A.A. du Plessis, and E.J.N. Looijenga. Topological stability of smooth mappings. Lecture Notes in Mathematics, Vol. 552. Springer-Verlag, Berlin-New York, 1976.

  36. 36.

    H.H. Goldstine and J. von Neumann. Numerical inverting matrices of high order, II. Proceedings of the Amer. Math. Soc., 2:188–202, 1951.

    MathSciNet  MATH  Article  Google Scholar 

  37. 37.

    D.Yu. Grigoriev. Complexity of deciding Tarski algebra. Journal of Symbolic Computation, 5:65–108, 1988.

    MathSciNet  MATH  Article  Google Scholar 

  38. 38.

    D.Yu. Grigoriev and N.N. Vorobjov. Solving systems of polynomial inequalities in subexponential time. Journal of Symbolic Computation, 5:37–64, 1988.

    MathSciNet  MATH  Article  Google Scholar 

  39. 39.

    D.Yu. Grigoriev and N.N. Vorobjov. Counting connected components of a semialgebraic set in subexponential time. Computational Complexity, 2:133–186, 1992.

    MathSciNet  MATH  Article  Google Scholar 

  40. 40.

    A. Hatcher. Algebraic topology. Cambridge University Press, Cambridge, 2002.

    MATH  Google Scholar 

  41. 41.

    J. Heintz, M.-F. Roy, and P. Solerno. Single exponential path finding in semi-algebraic sets II: The general case. In C.L. Bajaj, editor, Algebraic Geometry and its Applications, pages 449–465. Springer-Verlag, 1994.

  42. 42.

    R.E. Hodel. An Introduction to Mathematical Logic. Dover Publications Inc., Mineola, New York, 2013.

    Google Scholar 

  43. 43.

    P. Koiran. The real dimension problem is \({\rm NP}_{\mathbb{R}}\)-complete. J. Complexity, 15:227–238, 1999.

    MathSciNet  MATH  Article  Google Scholar 

  44. 44.

    E. Kostlan. Complexity theory of numerical linear algebra. J. of Computational and Applied Mathematics, 22:219–230, 1988.

    MathSciNet  MATH  Article  Google Scholar 

  45. 45.

    P. Lairez. A deterministic algorithm to compute approximate roots of polynomial systems in polynomial average time. Found. Comput. Math., 17(5):1265–1292, 2017.

    MathSciNet  MATH  Article  Google Scholar 

  46. 46.

    J. Mather. Notes on topological stability. Bull. Amer. Math. Soc. (N.S.), 49(4):475–506, 2012.

    MathSciNet  MATH  Article  Google Scholar 

  47. 47.

    J. von Neumann and H.H. Goldstine. Numerical inverting matrices of high order. Bull. Amer. Math. Soc., 53:1021–1099, 1947.

    MathSciNet  MATH  Article  Google Scholar 

  48. 48.

    P. Niyogi, S. Smale, and S. Weinberger. Finding the homology of submanifolds with high confidence from random samples. Discrete Comput. Geom., 39(1-3):419–441, 2008.

    MathSciNet  MATH  Article  Google Scholar 

  49. 49.

    J. Renegar. On the computational complexity and geometry of the first-order theory of the reals. Part I. Journal of Symbolic Computation, 13:255–299, 1992.

    MathSciNet  MATH  Article  Google Scholar 

  50. 50.

    D.J.S. Robinson. A course in the theory of groups, volume 80 of Graduate Texts in Mathematics. Springer-Verlag, New York, second edition, 1996.

  51. 51.

    J.J. Rotman. An introduction to algebraic topology, volume 119 of Graduate Texts in Mathematics. Springer-Verlag, New York, 1988.

  52. 52.

    J.J. Rotman. An introduction to homological algebra. Universitext. Springer, New York, second edition, 2009.

    MATH  Book  Google Scholar 

  53. 53.

    J.T. Schwartz and M. Sharir. A survey of motion planning and related geometric algorithms. Artificial Intelligence, 37:157–169, 1988.

    MathSciNet  MATH  Article  Google Scholar 

  54. 54.

    M. Shub and S. Smale. Complexity of Bézout’s Theorem I: geometric aspects. Journal of the Amer. Math. Soc., 6:459–501, 1993.

    MathSciNet  MATH  Google Scholar 

  55. 55.

    M. Shub and S. Smale. Complexity of Bézout’s Theorem II: volumes and probabilities. In F. Eyssette and A. Galligo, editors, Computational Algebraic Geometry, volume 109 of Progress in Mathematics, pages 267–285. Birkhäuser, 1993.

  56. 56.

    M. Shub and S. Smale. Complexity of Bézout’s Theorem III: condition number and packing. Journal of Complexity, 9:4–14, 1993.

    MathSciNet  MATH  Article  Google Scholar 

  57. 57.

    M. Shub and S. Smale. Complexity of Bézout’s Theorem V: polynomial time. Theoret. Comp. Sci., 133:141–164, 1994.

    MATH  Article  Google Scholar 

  58. 58.

    M. Shub and S. Smale. Complexity of Bézout’s Theorem IV: probability of success; extensions. SIAM J. of Numer. Anal., 33:128–148, 1996.

    MATH  Article  Google Scholar 

  59. 59.

    S. Smale. Complexity theory and numerical analysis. In A. Iserles, editor, Acta Numerica, pages 523–551. Cambridge University Press, 1997.

  60. 60.

    S. Smale. Mathematical problems for the next century. Mathematical Intelligencer, 20:7–15, 1998.

    MathSciNet  MATH  Article  Google Scholar 

  61. 61.

    E. Spanier. Algebraic Topology. McGraw-Hill, New York, 1966.

    MATH  Google Scholar 

  62. 62.

    A. Tarski. A decision method for elementary algebra and geometry. University of California Press, Berkeley and Los Angeles, Calif., 1951. 2nd ed.

  63. 63.

    R. Thom. Ensembles et morphismes stratifiés. Bull. Amer. Math. Soc., 75:240–284, 1969.

    MathSciNet  MATH  Article  Google Scholar 

  64. 64.

    A.M. Turing. Rounding-off errors in matrix processes. Quart. J. Mech. Appl. Math., 1:287–308, 1948.

    MathSciNet  MATH  Article  Google Scholar 

  65. 65.

    H.R. Wüthrich. Ein Entscheidungsverfahren für die Theorie der reell-abgeschlossenen Körper. In E. Specker and V. Strassen, editors, Komplexität von Entscheidungsproblemen, volume 43 of Lect. Notes in Comp. Sci., pages 138–162. Springer-Verlag, 1976.

  66. 66.

    G.M. Ziegler. Lectures on Polytopes. Graduate Texts in Mathematics. Springer New York, 2012.

Download references

Acknowledgements

We are grateful to Saugata Basu, Pierre Lairez, and Nicolai Vorobjov for helpful discussions.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Felipe Cucker.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This work was supported by the Einstein Foundation, Berlin.

Peter Bürgisser: Partially funded by DFG research Grant BU 1371/2-2.

Felipe Cucker: Partially supported by a GRF grant from the Research Grants Council of the Hong Kong SAR (Project Number CityU 11302418).

Shmuel Weinberger.

A On the Smoothness Assumption in Thom’s First Isotopy Lemma

A On the Smoothness Assumption in Thom’s First Isotopy Lemma

We begin observing that we can define Whitney stratifications of any subset of a manifold in the same manner we define Whitney stratifications of the manifold itself.

The following lemma will be instrumental in our proof.

Lemma A.1

Let \(\mathcal {M}\) be a smooth manifold and \(\mathcal {S}\) be a locally finite partition of a locally closed subset \(\Omega \subset \mathcal {M}\). Then:

  1. 1.

    Let \(\mathcal {S}^c\) be the partition whose elements are the connected components of the elements in \(\mathcal {S}\). If \(\mathcal {S}\) is a Whitney stratification, then so is \(\mathcal {S}^c\).

  2. 2.

    If \(\mathcal {S}\) is a Whitney stratification with connected strata, then it satisfies the boundary condition:

    $$\begin{aligned} \text{ for } \sigma ,\varsigma \in \mathcal {S}, \text{ if } \varsigma \cap \overline{\sigma }\ne \varnothing , \text{ then } \varsigma \subseteq \overline{\sigma }. \end{aligned}$$
    (BC)
  3. 3.

    Let \(\mathcal {S}\) satisfy the boundary condition (BC). Then, \(\mathcal {S}\) is a Whitney stratification if and only if for all \(\sigma \in \mathcal {S}\),

    $$\begin{aligned} \mathcal {S}_{|\overline{\sigma }}:=\{\varsigma \in \mathcal {S}\mid \varsigma \subseteq \overline{\sigma } \} \end{aligned}$$

    is a Whitney stratification of \(\overline{\sigma }\).

  4. 4.

    Let \(\mathcal {S}\) be a Whitney stratification, \(\{\sigma _i\}_{i\in I}\subseteq \mathcal {S}\) a family of strata of the same dimension and \(\mathcal {S}'\) the partition obtained from \(\mathcal {S}\) by replacing the \(\sigma _i\) by its union. Then, \(\mathcal {S}'\) is a Whitney stratification.

Proof

Parts 1 and 2 are [35, Ch. II, Theorem 5.6 and Corollary 5.7], respectively. For part 3, the fact that \(\mathcal {S}\) satisfies (BC) implies that \(\mathcal {S}_{|\overline{\sigma }}\) is a partition of \(\overline{\sigma }\) whose elements are elements in \(\mathcal {S}\). As a subset of a Whitney stratification is a Whitney stratification itself we have shown the “only if” part. We next show the converse. For every \(\sigma \in \mathcal {S}\), the fact that \(\mathcal {S}_{|\overline{\sigma }}\) is a Whitney stratification implies that \(\sigma \in \mathcal {S}_{|\overline{\sigma }}\) is a locally closed smooth submanifold. Next note that Whitney’s condition b needs to be checked only for pairs \((\sigma ,\varsigma )\in \mathcal {S}^2\) such that \(\varsigma \cap \overline{\sigma }\ne \varnothing \). But the fact that \(\mathcal {S}\) satisfies (BC) implies that, for any such pair, \(\sigma ,\varsigma \in \mathcal {S}_{|\overline{\sigma }}\) and therefore, it satisfies condition b because, by hypothesis, \(\mathcal {S}_{|\overline{\sigma }}\) is a Whitney stratification.

We finally prove 4. By the local character of Definition 4.1 of Whitney stratification, it is enough to check the conditions in this definition in some open neighborhood \(U_x\) around each point \(x\in \Omega \). Since \(\mathcal {S}\) is locally finite, we can pick each \(U_x\) such that \(\mathcal {S}_{|U_x}\) is finite. Hence, without loss of generality, we can assume that I is finite.

For all \(i\ne j\in I\) we have \(\sigma _i\cap \overline{\sigma _j}=\varnothing \). Otherwise, by [35, Ch. I, (1.1)], we would have \(\dim \sigma _i< \dim \sigma _j\), contradicting our hypothesis. Hence, for all \(i\in I\), there is an open set \(U_i\) such that \(\sigma _i\subseteq U_i\) and, for all \(j\ne i\), \(U_i\cap \overline{\sigma _j}=\varnothing \). It follows that \(\cup \sigma _i\) is a locally closed smooth manifold. The verification of the conditions in Definition 4.1 for \(\mathcal {S}'\) is now straightforward. \(\square \)

To prove Theorem 4.4, we will rely on the following version of Thom’s First Isotopy Lemma which is the one in [35, Ch. II, Theorem 5.2].

Theorem A.2

Let \(\mathcal {M}\) be a smooth manifold and \(\Omega \subseteq \mathcal {M}\) a locally closed subset with a Whitney stratification \(\mathcal {S}\) and let \(\alpha :\mathcal {M}\rightarrow \mathbb {R}^k\) be a smooth proper map such that:

  • for each stratum \(\sigma \in \mathcal {S}\), \(\alpha _{|\sigma }:\sigma \rightarrow \mathbb {R}^k\) is surjective,

  • for each stratum \(\sigma \in \mathcal {S}\), \(\alpha _{|\sigma }:\sigma \rightarrow \mathbb {R}^k\) is a smooth submersion.

Then, \(\alpha _{|X}\) is a trivial fiber bundle. \(\square \)

To deduce Theorem 4.4 from this result, we will employ graphs of maps. This will allow us to transform our not necessarily smooth map into a smooth one, as it will be simply a projection.

Let A and B be smooth manifolds. Recall that the graph of a function \(\varphi :A\rightarrow B\) is the set

$$\begin{aligned} \Gamma _\varphi :=\{(a,b)\in A\times B\mid \varphi (a)=b\}. \end{aligned}$$

Associated with the graph, we have the functions \(i_\varphi :A\rightarrow \Gamma _\varphi \), given by \(a\mapsto (a,\varphi (a))\), and \(\pi :A\times B\rightarrow B\), given by \((a,b)\mapsto b\). Clearly, \(\varphi =\pi \circ i_\varphi \). Also, it is easy to see, if \(\varphi \) is a continuous map, then \(\Gamma _\varphi \) is a closed subset of \(A\times B\) and \(i_\varphi \) is a homeomorphism between A and \(\Gamma _\varphi \). Finally, if \(\varphi \) is a smooth map, then \(\Gamma _\varphi \) is a closed smooth submanifold of \(A\times B\) and \(i_\varphi \) is a diffeomorphism between A and \(\Gamma _\varphi \). Given a subset \(X\subseteq A\), we will consider

$$\begin{aligned} \Gamma _\varphi (X):=\Gamma _{\varphi _{|X}}= \{(a,b)\in \Gamma _\varphi \mid a\in X\}= \{(a,\varphi (a))\mid a\in X\}. \end{aligned}$$

It is again clear that if \(\varphi \) is continuous and X is a locally closed subset of A, then \(\Gamma _\varphi (X)\) is a locally closed subset of \(A\times B\). Moreover, if \(\varphi \) is smooth and X is a locally closed smooth submanifold of A, then \(\Gamma _\varphi (X)\) is a locally closed smooth submanifold of \(A\times B\).

Proof of Theorem 4.4

Consider the graph \(\Gamma _{\alpha }\) of \(\alpha \). Although not necessarily a manifold (as \(\alpha \) may be non-smooth), it is a locally closed subset of \(\mathcal {M}\times \mathbb {R}^k\). Next consider the partition of \(\Gamma _\alpha \) given by

$$\begin{aligned} \Gamma _\alpha (\mathcal {S}):= \{\Gamma _\alpha (\sigma )\mid \sigma \in \mathcal {S}\} \end{aligned}$$

and its associated partition \(\Gamma ^c_\alpha (\mathcal {S})\) as defined in Lemma A.1(1).

We claim that \(\Gamma ^c_\alpha (\mathcal {S})\) is a Whitney stratification of \(\Gamma _\alpha \).

To prove the claim we first observe that \(\Gamma ^c_\alpha (\mathcal {S})=\Gamma _\alpha (\mathcal {S}^c)\). As, by Lemma A.1(1), \(\mathcal {S}^c\) is a Whitney stratification and, by construction, has connected strata, Lemma A.1(2) shows that it satisfies the boundary condition (BC). It follows that \(\Gamma ^c_\alpha (\mathcal {S})\) satisfies (BC) as well.

Let \(\sigma \in \mathcal {S}^c\) and \(\sigma '\in \mathcal {S}\) such that \(\sigma \subseteq \sigma '\). By the first hypothesis in our statement, there is an open neighborhood U of \(\overline{\sigma '}\supseteq \overline{\sigma }\) and a smooth map \(\varphi :U\rightarrow \mathbb {R}^k\) such that \(\alpha _{|\sigma '}=\varphi \). Clearly, \(\alpha _{|\sigma }=\varphi \) as well. This implies that \(\Gamma _\alpha (\mathcal {S}^c_{|\overline{\sigma }}) =\Gamma _\varphi (\mathcal {S}^c_{|\overline{\sigma }})\) and \(\Gamma _\alpha (\overline{\sigma }) =\Gamma _\varphi (\overline{\sigma })\). Since \(\varphi \) is smooth, \(\Gamma _\varphi \) is a locally closed smooth submanifold and \(i_\varphi :U\rightarrow \Gamma _\varphi \) is a diffeomorphism mapping the Whitney stratification \(\mathcal {S}^c_{|\overline{\sigma }}\) to \(\Gamma ^c_\alpha (\mathcal {S}_{|\overline{\sigma }})\) and the closed set \(\overline{\sigma }\) to \(\Gamma _\alpha (\overline{\sigma })\). Hence, by [35, Ch. I, (1.4)], \(\Gamma ^c_\alpha (\mathcal {S}_{|\overline{\sigma }})\) is a Whitney stratification of \(\Gamma _\alpha (\overline{\sigma })\).

As this happens for all strata \(\Gamma ^c_\alpha (\overline{\sigma })\) of the partition \(\Gamma ^c_\alpha (\mathcal {S})\) we may apply Lemma A.1(3) to deduce that \(\Gamma ^c_\alpha (\mathcal {S})\) is a Whitney stratification. We finally apply Lemma A.1(4) (several times for each dimension) to deduce that \(\Gamma _\alpha (\mathcal {S})\) itself is a Whitney stratification. This proves the claim.

Since \(\Gamma _\alpha (\mathcal {S})\) is a Whitney stratification of \(\Gamma _\alpha \) the map \(i_{\alpha }\) restricts to a diffeomorphism between \(\sigma \) and \(\Gamma _\alpha (\sigma )\), for all \(\sigma \in \mathcal {S}\). In addition, as \(\alpha =\pi \circ i_\alpha \), we have \(\alpha _{|\sigma }=\pi _{|\Gamma _\alpha (\sigma )}\circ (i_\alpha )_{|\sigma }\) and, hence, as \((i_\alpha )_{|\sigma }\) is a diffeomorphism, \(\pi _{|\Gamma _\alpha (\sigma )}\) is surjective if and only if \(\alpha _{|\sigma }\) is so, and \(\pi _{|\Gamma _\alpha (\sigma )}\) is a smooth submersion if and only if so is \(\alpha _{|\sigma }\). In summary, the last two hypotheses of our statement imply the hypothesis of Theorem A.2, and consequently, that \(\pi _{|\Gamma _\alpha }\) is a trivial bundle.

We can now conclude because a trivialization \(h:\Gamma _\alpha \rightarrow F\times \mathbb {R}^k\) of \(\pi _{|\Gamma _\alpha }\) induces the trivialization \(h\circ i_\alpha \) of \(\alpha :\mathcal {M}\rightarrow \mathbb {R}^k\). \(\square \)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Bürgisser, P., Cucker, F. & Tonelli-Cueto, J. Computing the Homology of Semialgebraic Sets. I: Lax Formulas. Found Comput Math 20, 71–118 (2020). https://doi.org/10.1007/s10208-019-09418-y

Download citation

Keywords

  • Homology groups
  • Weak complexity
  • Numerical algorithms

Mathematics Subject Classification

  • 14P10
  • 65D18
  • 65Y20
  • 68Q25