Improved bounds on absolute positiveness of multivariate polynomials

Abstract A multivariate polynomial F ( x 1 , x 2 , … , x n ) is said to be absolutely positive from a real number B if F and all its partial derivatives are non-negative for x 1 , x 2 , … , x n ≥ B . One of the well known bounds on absolute positiveness in the literature is due to Hong. His bound is dependent on the largest number in a certain sequence of radicals defined using the absolute value of the coefficients of the polynomial. In the univariate setting, a bound due to Lagrange considers the first and the second largest numbers in the same radical sequence and is shown by Collins to be better than Hong's bound. In the 1930's, Westerfield had proposed a bound that considers every value in the same radical sequence and improves on Lagrange's bound. In this paper, we provide a hierarchy of bounds generalizing Westerfield's bound to the multivariate setting. One of the bounds in this hierarchy is a generalization of Lagrange's bound. All the bounds are strict quantitative improvements over Hong's bound. We also give an algorithm to compute the multivariate Lagrange bound. The running time of this algorithm matches the running time of the best known algorithm to compute Hong's bound. The algorithm uses the range tree data structure to implement orthogonal range querying. Relying on a result of Fredman (1984), we show that the efficiency of this algorithm cannot be improved by using any other data structure for orthogonal range querying.

[1]  J B Kiostelikis,et al.  Bounds for positive roots of polynomials , 1986 .

[2]  Maurice Mignotte,et al.  Improvements of Lagrange's bound for polynomial roots , 2017, J. Symb. Comput..

[3]  Ronald L. Rivest,et al.  Introduction to Algorithms , 1990 .

[4]  Hoon Hong,et al.  Bounds for Absolute Positiveness of Multivariate Polynomials , 1998, J. Symb. Comput..

[5]  Vikram Sharma,et al.  A Lower Bound for Computing Lagrange's Real Root Bound , 2016, CASC.

[6]  Vikram Sharma,et al.  Stronger Tradeoffs for Orthogonal Range Querying in the Semigroup Model , 2018, FSTTCS.

[7]  George E. Collins.  Krandick's proof of Lagrange's real root bound claim , 2015, J. Symb. Comput..

[8]  Vikram Sharma,et al.  Improved Bounds on Absolute Positiveness of Multivariate Polynomials , 2017, ISSAC.

[9]  Dirk van Dalen,et al.  How the Mathematical Objects Determine the Mathematical Principles , 2005, J. Univers. Comput. Sci..

[10]  Kurt Mehlhorn,et al.  Faster algorithms for computing Hong's bound on absolute positiveness , 2010, J. Symb. Comput..

[11]  Doru ŞtefăźNescu.  A New Polynomial Bound and Its Efficiency , 2015, CASC 2015.

[12]  Hoon Hong,et al.  Testing Positiveness of Polynomials , 1998, Journal of Automated Reasoning.

[13]  Mark de Berg,et al.  Computational geometry: algorithms and applications , 1997 .

[14]  Prashant Batra,et al.  On the Quality of Some Root-Bounds , 2015, MACIS.

[15]  E. C. Westerfield.  A New Bound for the Zeros of Polynomials , 1933 .

[16]  Vikram Sharma,et al.  Bounds on absolute positiveness of multivariate polynomials , 2010, J. Symb. Comput..

[17]  G.S. Brodal,et al.  Dynamic planar convex hull , 2002, The 43rd Annual IEEE Symposium on Foundations of Computer Science, 2002. Proceedings..

[18]  Michael L. Fredman,et al.  A Lower Bound on the Complexity of Orthogonal Range Queries , 1981, JACM.

[19]  Kurt Mehlhorn,et al.  Data Structures and Algorithms 3: Multi-dimensional Searching and Computational Geometry , 2012, EATCS Monographs on Theoretical Computer Science.

[20]  J. Lagrange.  De la résolution des équations numériques de tous les degrés , 1970 .

[21]  Santosh S. Vempala,et al.  Accelerated Newton Iteration for Roots of Black Box Polynomials , 2016, 2016 IEEE 57th Annual Symposium on Foundations of Computer Science (FOCS).

[22]  Alkiviadis G. Akritas,et al.  Implementations of a New Theorem for Computing Bounds for Positive Roots of Polynomials , 2006, Computing.