Abstract
This paper investigates the cost of solving systems of sparse polynomial equations by homotopy continuation. First, a space of systems of n-variate polynomial equations is specified through n monomial bases. The natural locus for the roots of those systems is known to be a certain toric variety. This variety is a compactification of \((\mathbb {C}\setminus \{0\})^n\), dependent on the monomial bases. A toric Newton operator is defined on that toric variety. Smale’s alpha theory is generalized to provide criteria of quadratic convergence. Two condition numbers are defined, and a higher derivative estimate is obtained in this setting. The Newton operator and related condition numbers turn out to be invariant through a group action related to the momentum map. A homotopy algorithm is given and is proved to terminate after a number of Newton steps which is linear on the condition length of the lifted homotopy path. This generalizes a result from Shub (Found Comput Math 9(2):171–178, 2009. https://doi.org/10.1007/s10208-007-9017-6).
Similar content being viewed by others
References
Allgower, Eugene L. and Kurt Georg. 1993. Continuation and path following, Acta numerica, 1993, Acta Numer., Cambridge Univ. Press, Cambridge, pp. 1–64.
Armentano, Diego, Carlos Beltrán, Peter Bürgisser, Felipe Cucker, and Michael Shub. 2016. Condition Length and Complexity for the Solution of Polynomial Systems, Found. Comput. Math., 16, no. 6, 1401–1422, https://doi.org/10.1007/s10208-016-9309-9.
Beltrán, Carlos. 2011. A continuation method to solve polynomial systems and its complexity, Numer. Math. 117, no. 1, 89–113, https://doi.org/10.1007/s00211-010-0334-3.
Beltrán, Carlos, Jean-Pierre Dedieu, Gregorio Malajovich, and Mike Shub. 2009 Convexity properties of the condition number, SIAM J. Matrix Anal. Appl. 31,no. 3, 1491–1506, https://doi.org/10.1137/080718681.
Beltrán, Carlos, Jean-Pierre Dedieu, Gregorio Malajovich, and Mike Shub. 2012. Convexity properties of the condition number II, SIAM J. Matrix Anal. Appl. 33, no. 3, 905–939, https://doi.org/10.1137/100808885.
Beltrán, Carlos, Anton Leykin. 2013. Robust certified numerical homotopy tracking, Found. Comput. Math. 13, no. 2, 253–295, https://doi.org/10.1007/s10208-013-9143-2.
Beltrán, Carlos, Luis Miguel Pardo. 2009. Smale’s 17th problem: average polynomial time to compute affine and projective solutions, J. Amer. Math. Soc. 22, no. 2, 363–385, https://doi.org/10.1090/S0894-0347-08-00630-9.
Beltrán, Carlos, Luis Miguel Pardo. 2011. Fast linear homotopy to find approximate zeros of polynomial systems, Found. Comput. Math. 11, no. 1, 95–129, https://doi.org/10.1007/s10208-010-9078-9.
Beltrán, Carlos and Michael Shub. 2009. Complexity of Bezout’s theorem. VII. Distance estimates in the condition metric, Found. Comput. Math. 9, no. 2, 179–195, https://doi.org/10.1007/s10208-007-9018-5.
Bernstein, D. N., A. G. Kušnirenko, and A. G. Hovanskiĭ. 1976. Newton polyhedra, Uspehi Mat. Nauk 31, no. 3(189), 201–202 (Russian).
Blum, Lenore, Felipe Cucker, Michael Shub, and Steve Smale. 1998. Complexity and real computation, Springer-Verlag, New York. With a foreword by Richard M. Karp.
Boito, Paola and Jean-Pierre Dedieu. 2010. The condition metric in the space of rectangular full rank matrices, SIAM J. Matrix Anal. Appl. 31, no. 5, 2580–2602, https://doi.org/10.1137/08073874X.
Bürgisser, Peter and Felipe Cucker. 2011. On a problem posed by Steve Smale, Ann. of Math. (2) 174, no. 3, 1785–1836, https://doi.org/10.4007/annals.2011.174.3.8.
Criado del Rey, Juan. TA. Condition metrics in the three classical spaces, arxiv:1501.04456
Dedieu, Jean-Pierre, Gregorio Malajovich and Michael Shub. 2013. Adaptive step-size selection for homotopy methods to solve polynomial equations, IMA J. Numer. Anal. 33, 1–29, https://doi.org/10.1093/imanum/drs007.
Dedieu, Jean-Pierre, Pierre Priouret and Gregorio Malajovich. 2003. Newton’s method on Riemannian manifolds: convariant alpha theory, IMA J. Numer. Anal. 23, no. 3, 395–419, https://doi.org/10.1093/imanum/23.3.395.
Hauenstein, Jonathan D. and Alan C. Liddell Jr. 2016. Certified predictor-corrector tracking for Newton homotopies, J. Symbolic Comput. 74, 239–254, https://doi.org/10.1016/j.jsc.2015.07.001.
Jensen, Anders. TA. Tropical Homotopy Continuation, arxiv:1601.02818
Knuth, Donald E. 2005. The art of computer programming. Vol. 4, Fasc. 3. Addison-Wesley, Upper Saddle River, NJ. Generating all combinations and partitions.
Lairez, Pierre. 2017. A deterministic algorithm to compute approximate roots of polynomial systems in polynomial average time, Foundations of Computational Mathematics 17, no. 5, 1265-1292 https://doi.org/10.1007/s10208-016-9319-7.
Li, Chong and Wang, Jinhua. 2008. Newton’s method for sections on Riemannian manifolds: generalized covariant \(\alpha \) -theory, J. Complexity 24, no. 3, 423–451, https://doi.org/10.1016/j.jco.2007.12.003.
Malajovich, Gregorio. 2011. Nonlinear equations, Publicações Matemáticas do IMPA, \(28^{\rm o}\) Colóquio Brasileiro de Matemática., Instituto Nacional de Matemática Pura e Aplicada (IMPA), Rio de Janeiro. Available at http://www.labma.ufrj/~gregorio.
Malajovich, Gregorio. 2013a. On the expected number of zeros of nonlinear equations, Found. Comput. Math. 13, no. 6, 867–884, https://doi.org/10.1007/s10208-013-9171-y.
Malajovich, Gregorio. 2013b. Newton iteration, conditioning and zero counting, Recent advances in real complexity and computation, Contemp. Math., vol. 604, Amer. Math. Soc., Providence, RI, , pp. 151–185, https://doi.org/10.1090/conm/604/12072.
Malajovich, Gregorio. 2017. Computing mixed volume and all mixed cells in quermassintegral time, Found. Comput. Math., 17, no. 5, 1293-1334, https://doi.org/10.1007/s10208-016-9320-1.
Malajovich, Gregorio and J. Maurice Rojas. 2004. High probability analysis of the condition number of sparse polynomial systems, Theoret. Comput. Sci. 315, no. 2-3, 524–555, https://doi.org/10.1016/j.tcs.2004.01.006.
Maxima. 2014. Maxima, a Computer Algebra System, Version 5.34.1, Available at http://maxima.sourceforge.net, last update: 2014.09.08.
Morgan, Alexander. 2009. Solving polynomial systems using continuation for engineering and scientific problems, Classics in Applied Mathematics, vol. 57, Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA. Reprint of the 1987 original; Pages 304–534: computer programs section, also available as a separate file online.
Shub, Michael. 1993. Some remarks on Bezout’s theorem and complexity theory, From Topology to Computation: Proceedings of the Smalefest (Berkeley, CA, 1990), Springer, New York, pp. 443-455.
Shub, Michael. 2009. Complexity of Bezout’s theorem. VI. Geodesics in the condition (number) metric, Found. Comput. Math. 9, no. 2, 171–178, https://doi.org/10.1007/s10208-007-9017-6.
Shub, Michael and Steve Smale. 1993a. Complexity of Bézout’s theorem. I. Geometric aspects, J. Amer. Math. Soc. 6, no. 2, 459–501, https://doi.org/10.2307/2152805.
Shub, Michael and Steve Smale. 1993b. Complexity of Bezout’s theorem. II. Volumes and probabilities, Computational algebraic geometry, (Nice, 1992), Progr. Math., vol. 109, Birkhäuser Boston, Boston, MA, pp. 267–285.
Shub, Michael and Steve Smale. 1993c. Complexity of Bezout’s theorem. III. Condition number and packing, J. Complexity, 9, no. 1, 4–14, https://doi.org/10.1006/jcom.1993.1002. Festschrift for Joseph F. Traub, Part I.
Shub, Michael and Steve Smale. 1994. Complexity of Bezout’s theorem. V. Polynomial time, Theoret. Comput. Sci. 133, no. 1, 141–164, https://doi.org/10.1016/0304-3975(94)90122-8. Selected papers of the Workshop on Continuous Algorithms and Complexity (Barcelona, 1993).
Shub, Michael and Steve Smale. 1996 Complexity of Bezout’s theorem. IV. Probability of success; extensions, SIAM J. Numer. Anal. 33, no. 1, 128–148, https://doi.org/10.1137/0733008.
Sloane, N.J.A. (ed.) 2016. The On-Line Encyclopedia of Integer Sequences.
Smale, Steve. 1986. Newton’s method estimates from data at one point, The merging of disciplines: new directions in pure, applied, and computational mathematics (Laramie, Wyo., 1985), Springer, New York, pp. 185–196.
Smale, Steve. 1998. Mathematical problems for the next century, Math. Intelligencer 20, no. 2, 7–15, https://doi.org/10.1007/BF03025291.
Wang, Xing Hua. 1993. Some results relevant to Smale’s reports, From Topology to Computation: Proceedings of the Smalefest (Berkeley, CA, 1990), Springer, New York, pp. 456–465.
Acknowledgements
The author would like to thank Carlos Beltrán, Bernardo Freitas Paulo da Costa, Felipe Bottega Diniz and two anonymous referees for their suggestions and improvements.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Felipe Cucker.
Part of these results were obtained while visiting the Simons Institute for the Theory of Computing in the University of California at Berkeley. This visit was funded by CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Brazil. Proc. BEX 2388/14-6). This research was also funded by CNPq Grants 441678/2014-9 and 306673/2013-4.
Appendix A: Proof of Lemma 2.5.3
Appendix A: Proof of Lemma 2.5.3
We start with a real version of Lemma 2.5.3. This will be used to recover the complex version. The notation \(\langle \cdot . \cdot \rangle \) stands for the canonical Hermitian inner product in \(\mathbb {C}^n\), and \(\langle \cdot . \cdot \rangle _{\mathbb {R}^n}\) is the real canonical inner product. Identifying \(\mathbb {C}^n\) to \(\mathbb {R}^{2n}\) we can write
Since the same norm arises from those two inner products, we use the notation \(\Vert \cdot \Vert \) for it. Here is the real lemma:
Lemma A.1
Suppose that \(\mathbf {x}, \mathbf {y}, \varvec{\zeta }\in \mathbb {R}^{n+1}\) with \(\varvec{\zeta }-\mathbf {x} \perp \mathbf {x}\), \(\mathbf {y}-\mathbf {x} \perp \mathbf {x}\) and \(\Vert \mathbf {y}-\varvec{\zeta }\Vert \le \Vert \mathbf {x}-\varvec{\zeta }\Vert \). Then,
where \(\pi _{\mathbb {R}}(\mathbf {y}) = \frac{\Vert \varvec{\zeta }\Vert ^2}{\langle \mathbf {y},\varvec{\zeta }\rangle _{\mathbb {R}^{n+1}}} \mathbf {y} \) is the radial projection onto the real affine plane \(\varvec{\zeta }+\varvec{\zeta }^{\perp }\).
Proof
Rescaling the three vectors \(\mathbf {x}, \mathbf {y}\) and \(\varvec{\zeta }\) simultaneously we can assume that \(\Vert \mathbf {x}\Vert =1\). Then we can choose an orthonormal basis \((\mathbf e_0, \dots , \mathbf e_n)\) so that \(\mathbf {x} = \mathbf e_0\), \(\varvec{\zeta }\) is in the span of \(\mathbf e_0\) and \(\mathbf e_1\) and y is in the span of \(\mathbf e_0, \mathbf e_1\) and \(\mathbf e_2\). In coordinates,
We can further assume that \(t \ge 0\) and \(r \ge 0\). Squaring both sides of the hypothesis \(\Vert \mathbf {y}-\varvec{\zeta }\Vert \le \Vert \mathbf {x}-\varvec{\zeta }\Vert \) we obtain
that is
which implies \(s \ge 0\).
We claim first that
We compute
To show in Eq. (15), we just need to verify that
Using the Maxima computer algebra system [27],
with
From the factorization above, K is negative if and only if \(A r^2 + B \ge 0\). Clearly, \(B \ge 0\). If \(A \ge 0\) we are done, so assume \(A<0\). Then multiplying both sides of (14) by A, one obtains
and
This shows (15). Also,
Taking square roots and combining with (15),
\(\square \)
Lemma 2.5.3Suppose that\(\mathbf {x}, \mathbf {y}, \varvec{\zeta }\in \mathbb {C}^{n+1}\)with\(\varvec{\zeta }-\mathbf {x} \perp \mathbf {x}\), \(\mathbf {y}-\mathbf {x} \perp \mathbf {x}\)and\(\Vert \mathbf {y}-\varvec{\zeta }\Vert \le \Vert \mathbf {x}-\varvec{\zeta }\Vert \). Then,
where\(\pi (\mathbf {y}) = { \frac{\Vert \varvec{\zeta }\Vert ^2}{\langle \mathbf {y},\varvec{\zeta }\rangle } \mathbf {y}} \)is the radial projection onto the affine plane\(\varvec{\zeta }+\varvec{\zeta }^{\perp }\).
Proof
We identify \(\mathbb {C}^{n+1}\) with \(\mathbb {R}^{2n+2}\) and claim that
Since complex orthogonal vectors are also real orthogonal, Eq. (16) and Lemma A.1 imply
To show (16) we choose coordinates so that
with \(c \ge 0\). A straightforward computation gives
We have
with equality if \(b=0\). This finishes the proof. \(\square \)
Rights and permissions
About this article
Cite this article
Malajovich, G. Complexity of Sparse Polynomial Solving: Homotopy on Toric Varieties and the Condition Metric. Found Comput Math 19, 1–53 (2019). https://doi.org/10.1007/s10208-018-9375-2
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10208-018-9375-2