Abstract
In this article we analyze the global diffeomorphism property of polynomial maps \(F:\mathbb {R}^n\rightarrow \mathbb {R}^n\) by studying the properties of the Newton polytopes at infinity corresponding to the sum of squares polynomials \(\Vert F\Vert _2^2\). This allows us to identify a class of polynomial maps F for which their global diffeomorphism property on \(\mathbb {R}^n\) is equivalent to their Jacobian determinant \(\det JF\) vanishing nowhere on \(\mathbb {R}^n\). In other words, we identify a class of polynomial maps for which the Real Jacobian Conjecture, which was proven to be false in general, still holds.
Similar content being viewed by others
References
Avis, D., Fukuda, K.: A Pivoting algorithm for convex hulls and vertex enumeration of arrangements and polyhedra. Discret. Comput. Geom. 8, 295–313 (1992)
Bajbar, T., Stein, O.: Coercive polynomials and their Newton polytopes. SIAM J. Optim. 25(3), 1542–1570 (2015)
Banach, S., Mazur, B.: Über mehrdeutige stetige Abbildungen. Stud. Math. 5, 174–178 (1934)
Bass, H., Connell, E., Wright, D.: The Jacobian conjecture: reduction of degree and formal expansion of the inverse. Bull. Am. Math. Soc. 7, 287–330 (1982)
Białynicki-Birula, A., Rosenlicht, M.: Injective morphisms of real algebraic varieties. Proc. AMS 13, 200–203 (1962)
Bivià-Ausina, C.: Injectivity of real polynomial maps and Lojasiewicz exponents at infinity. Math. Z. 257(4), 745–767 (2007)
Bremner, D., Fukuda, K., Marzetta, A.: Primal-dual methods for vertex and facet enumeration. Discret. Comput. Geom. 20, 333–357 (1998)
Caccioppoli, R.: Sugli elements uniti delle transformazioni funzionali. Rc. Mat. Padova 3, 1–15 (1932)
Chen, Y., Dias, L.R.G., Takeuchi, K., Tibăr, M.: Invertible polynomial mappings via Newton non-degeneracy. Ann. l’Inst. Fourier 54(5), 1807–1822 (2014)
Chichilnisky, G.: Topology and invertible maps. Adv. Appl. Math. 21, 113–123 (1998)
De Angelis, V., Tuncel, S.: Handelman’s theorem on polynomials with positive multiples. In: Codes, Systems, and Graphical Models. The IMA Volumes in Mathematics and Its Applications, vol. 123, pp. 439–445 (2001)
De Marco, G., Gorni, G., Zampieri, G.: Global inversion of functions: an introduction. Nonlinear Differ. Equ. Appl. 1, 229–248 (1994)
Dinh, S.T., Ha, H.V., Pham, T.S., Thao, N.T.: Global Lojasiewicz-type inequality for non-degenerate polynomial maps. J. Math. Anal. Appl. 410(2), 541–560 (2014)
Drużkowski, L.: The Jacobian Conjecture: survey of some results. Banach Cent. Publ. 31(1), 163–171 (1995)
Fukui, T., Kurdyka, K., Paunescu, L.: Tame Nonsmooth inverse mapping theorems. SIAM J. Optim. 20(3), 1573–1590 (2010)
Gale, D., Nikaidô, H.: The Jacobian matrix and global univalence of mappings. Math. Ann. 159(2), 81–93 (1965)
Gao, S.: Absolute irreducibility of polynomials via Newton polytopes. J. Algebra 237, 501–520 (2001)
Gasiński, L., Papageorgiou, N.S.: Nonlinear Analysis. Series in Mathematical Analysis and Applications. Chapman & Hall/CRC Press, Taylor & Francis Group, Boca Raton (2005)
Gordon, W.B.: On the diffeomorphisms of Euclidean space. Am. Math. Mon. 79(7), 755–759 (1972)
Hadamard, J.S.: Sur les transformations planes. Comptes Rendus de l’Acadmie des Sciences Paris 142, 74 (1906)
Hadamard, J.S.: Sur les transformations ponctuelles. Bulletin de la Socit Mathmatique de France 34, 71–84 (1906)
Hadamard, J.S.: Sur les correspondances ponctuelles. Oeuvres, pp. 349–363. Editions du Centre Nationale de la Recherche Scientifique, Paris (1968)
Iliman, S., de Wolff, T.: Amoebas, nonnegative polynomials and sums of squares supported on circuits. Res. Math. Sci. 3(1), 1–35 (2016)
Kaveh, K., Khovanskii, A.G.: Algebraic equations and convex bodies. In: Itenberg, I., Jöricke, B., Passare, M. (eds.) Perspectives in Analysis, Geometry, and Topology, on the Occasion of the 60th Birthday of Oleg Viro, Progress in Mathematics, vol. 296, pp. 263–282. Birkhäuser Verlag Ag (2012)
Keller, O.: Ganze Cremona-Transformationen. Monatsh. Math. Phys. 47, 299–306 (1939)
Khovanskii, A., Esterov, A.: Elimination theory and Newton polytopes. Funct. Anal. Other Math. 2, 45–71 (2008)
Kouchnirenko, A.G.: Polyèdres de Newton et nombres de Milnor. Invent. Math. 32, 1–31 (1976)
Krantz, S.G.: Convex Analysis. CRC Press, Taylor & Francis Group, Boca Raton (2015)
Krantz, S.G., Parks, H.R.: The Implicit Function Theorem, History, Theory, and Applications. Birkhäuser Basel, New York (2003). doi:10.1007/978-1-4612-0059-8
Kurdyka, K., Rusek, K.: Polynomial-rational bijections of \(\mathbb{R}^n\). Proc. Am. Math. Soc. 102(4), 804–808 (1988)
Kurdyka, K., Rusek, K.: Surjectivity of certain injective semialgebraic transformations of \(\mathbb{R}^n\). Math. Z. 200, 141–148 (1988)
Kushnirenko, A.G.: Newton polytopes and the Bézout theorem. Funct. Anal. Appl. (translated from Russian) 10, 233–235 (1977)
Levy, M.P.: Sur les fonctions de ligne implicites. Bulletin de la Société Mathématique de France 48, 13–27 (1920)
Mas-Colell, A.: Homeomorphisms of compact, convex sets and the Jacobian matrix. SIAM J. Math. Anal. 10(6), 1105–1109 (1979)
Moh, T.T.: On the Jacobian conjecture and the configurations of roots. Journal für die reine und angewandte Mathematik 340, 140–212 (1983)
Netzer, T.: Stability of quadratic modules. Manuscr. Math. 129, 251–271 (2009)
Palais, R.S.: Natural operations on differential forms. Trans. Am. Math. Soc. 92, 121–141 (1959)
Pham, T.S.: On the topology of the Newton boundary at infinity. J. Math. Soc. Jpn. 60, 1065–1081 (2008)
Pinchuk, S.: A counterexample to the real Jacobian Conjecture. Math. Z. 217, 1–4 (1994)
Plastock, R.: Homeomorphisms between Banach spaces. Trans. Am. Math. Soc. 200, 169–183 (1974)
Rabier, P.J.: On global diffeomorphisms of Euclidean space. Nonlinear Anal. Theory Methods Appl. 21(12), 925–947 (1993)
Rabier, P.J.: Ehresmann fibrations and Palais–Smale conditions for morphisms of Finsler manifolds. Ann. Math. 146, 647–691 (1997)
Reznick, B.: Extremal PSD forms with few terms. Duke Math. J. 45, 363–374 (1978)
Ruzhansky, M., Sugimoto, M.: On global inversion of homogenous maps. Bull. Math. Sci. 5(1), 13–18 (2015)
Samuelson, P.A.: Prices of factors and goods in general equilibrium. Rev. Econ. Stud. 21(1), 1–20 (1953)
Sturmfels, B.: Polynomial equations and convex polytopes. Am. Math. Mon. 105, 907–922 (1998)
Thang, N.T.: Bifurcation set, M-tameness, asymptotic critical values and Newton polyhedrons. Kodai Math. J. 36, 77–90 (2013)
van den Essen, A.: To believe or not to believe: the Jacobian conjecture. Rendiconti del Seminario Matematico Universitè Politecnico di Torino 55(4), 283–290 (1997)
van den Essen, A.: Polynomial Automorphisms and the Jacobian Conjecture. Progress in Mathematics, vol. 190. Birkhäuser, Basel (2000)
Wang, S.: Jacobian criterion for separability. J. Algebra 21, 453–494 (1980)
Wright, D.: The Jacobian Conjecture: Ideal Membership Questions and Recent Advances. Contemporary Mathematics, vol. 369 (2005)
Yagzhev, A.V.: On Keller’s problem. Sib. Math. J. 21, 747–754 (1980)
Acknowledgements
The authors are grateful to Yu. Nesterov and V. Shikhman for pointing out the importance of the invariance of coercivity under linear transformations, and for other fruitful discussions on the subject of this article. The authors also wish to thank an anonymous referee for substantial remarks which significantly improved the quality of this manuscript.
Author information
Authors and Affiliations
Corresponding author
A Appendix
A Appendix
1.1 A.1 Proof of Lemma 1
For any \(\bar{v}\in {{\mathrm{vert}}}(P+P)\) there exists a vector \(a\in \mathbb {R}^n{\setminus }\{0\}\) such that \(\bar{v}\) is the unique optimal point of the problem
Let \(\bar{w}\in {{\mathrm{vert}}}(P)\) be an optimal point of the problem
Since
holds, the point \(2\bar{w}\in 2{{\mathrm{vert}}}(P)\) is an optimal point of (LP1). A a was chosen such that \(\bar{v}\) is the unique optimal point of (LP1), one obtains \(\bar{v}=2\bar{w}\) with \(\bar{w}\in {{\mathrm{vert}}}(P)\).
On the other hand, choose \(\bar{w}\in {{\mathrm{vert}}}(P)\) and put \(\bar{v}=2\bar{w}\). To show is \(\bar{v}\in {{\mathrm{vert}}}(P+P)\). Observe that there exists some \(a\in \mathbb {R}^n{\setminus }\{0\}\) such that \(\bar{w}\) is the unique optimal point of the problem (LP2). Using (A.1), the point \(\bar{v}\) is thus an optimal point of (LP1). Assume that \(\bar{v}\notin {{\mathrm{vert}}}(P+P)\) holds. Since (LP1) must possess a vertex solution, there exists an optimal point \(\bar{z}:=\bar{x}+\bar{y}\in P+P\) of (LP1) with \(\bar{v}\not =\bar{z}\). For the point \(\bar{u}:=\textstyle {\frac{1}{2}}(\bar{x}+\bar{y})\in P\) we obtain the identity
where the last equation holds since both \(\bar{z}\) and \(\bar{v}=2 \bar{w}\) are optimal for (LP1). The point \(\bar{u}\in P\) is thus an optimal point of the problem (LP2), and the uniqueness of \(\bar{w}\) implies \(\bar{w}=\bar{u}\). This leads to the contradiction \(\bar{v}=\bar{z}\), and thus the assertion \(\bar{v}\in {{\mathrm{vert}}}(P+P)\) follows. \(\square \)
1.2 A.2 Proof of Lemma 5
Let \(S_n\) denote the symmetric group on n elements, let \({{\mathrm{sign}}}(\sigma )\) denote the permutation sign of \(\sigma \in S_n\), and for some arbitrarily given \(x\in \mathbb {R}^n\) let the entries of JF(x) be denoted by \(a_{ij}\), \(i,j\in I\). Then the Leibniz formula for determinants yields
with
for all \(\sigma \in S_n\) and \(i\in I\). Interchanging multiplication and addition, and splitting the appearing products, further leads to
for all \(\sigma \in S_n\). In fact, in the above summation for any \(i\in I\) it is sufficient to choose \(\alpha ^i\in A(F_i)\) with \(\alpha ^i_{\sigma (i)}\ge 1\), since the existence of some \(j\in I\) with \(\alpha ^j_{\sigma (j)}=0\) means that the monomial \(x^{\alpha ^j}\) does not depend on the variable \(x_{\sigma (j)}\), resulting in
This shows
for all \(\sigma \in S_n\).
Next, for any \((\alpha ^1,\ldots ,\alpha ^n)\) in the above summation and any \(i\in I\) we have
and, since \(\sigma \) is a permutation,
We arrive at
for all \(\sigma \in S_n\), where the monomial
does not depend on \(\sigma \). Hence, we may write
In the latter outer summation it suffices to consider \(\alpha ^i\in A(F_i)\), \(i\in I\), with \(\sum _{i\in I}\alpha ^i\ge \mathbbm {1}\), since otherwise there would exist some \(j\in I\) with \(\alpha ^i_j=0\) for all \(i\in I\), resulting in \(\alpha ^{\sigma ^{-1}(j)}_j=0\) for any \(\sigma \in S_n\). However, then the inner summation would be taken over the empty set.
After introducing this restriction on the outer summation, we may drop the constraint \(\alpha ^i_{\sigma (i)}\ge 1\), \(i\in I\), in the inner summation since, for given \(\sigma \in S_n\), its violation leads to a vanishing product \(\prod _{i\in I}\alpha ^i_{\sigma (i)}\). Thus we have shown the assertion
where the final identity is due to the Leibniz formula for determinants. \(\square \)
Rights and permissions
About this article
Cite this article
Bajbar, T., Stein, O. On globally diffeomorphic polynomial maps via Newton polytopes and circuit numbers. Math. Z. 288, 915–933 (2018). https://doi.org/10.1007/s00209-017-1920-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00209-017-1920-1