AN ADDITIVE PROBLEM IN FINITE FIELDS WITH POWERS OF ELEMENTS OF LARGE MULTIPLICATIVE ORDER

Similar documents
Congruences involving product of intervals and sets with small multiplicative doubling modulo a prime

On the Fractional Parts of a n /n

arxiv: v2 [math.nt] 15 May 2013

arxiv: v2 [math.nt] 13 Dec 2013

ON POLYNOMIAL REPRESENTATION FUNCTIONS FOR MULTILINEAR FORMS. 1. Introduction

A LOWER BOUND FOR THE SIZE OF A MINKOWSKI SUM OF DILATES. 1. Introduction

CHARACTER-FREE APPROACH TO PROGRESSION-FREE SETS

RECOGNITION OF BEING FIBRED FOR COMPACT 3-MANIFOLDS

COUNTING NUMERICAL SEMIGROUPS BY GENUS AND SOME CASES OF A QUESTION OF WILF

Section II.1. Free Abelian Groups

Maximum exponent of boolean circulant matrices with constant number of nonzero entries in its generating vector

Solving a linear equation in a set of integers II

arxiv: v1 [math.nt] 15 Aug 2017

GENERALIZATIONS OF SOME ZERO-SUM THEOREMS. Sukumar Das Adhikari Harish-Chandra Research Institute, Chhatnag Road, Jhusi, Allahabad , INDIA

NEW UPPER BOUNDS FOR FINITE B h. Javier Cilleruelo

Weighted Sequences in Finite Cyclic Groups

Notes on Systems of Linear Congruences

FINITE FIELDS AND APPLICATIONS Additive Combinatorics in finite fields (3 lectures)

arxiv: v1 [math.na] 9 Feb 2013

On divisibility in definable groups

On symmetric square values of quadratic polynomials

On the solvability of groups with four class sizes.

Heights of characters and defect groups

Upper Bounds for Partitions into k-th Powers Elementary Methods

A locally convex topology and an inner product 1

RECOGNITION OF BEING FIBERED FOR COMPACT 3-MANIFOLDS

On the elliptic curve analogue of the sum-product problem

xy xyy 1 = ey 1 = y 1 i.e.

SUM-PRODUCT ESTIMATES APPLIED TO WARING S PROBLEM MOD P

MATH 324 Summer 2011 Elementary Number Theory. Notes on Mathematical Induction. Recall the following axiom for the set of integers.

A WHITNEY EXTENSION THEOREM FOR CONVEX FUNCTIONS OF CLASS C 1,ω.

Math 4310 Solutions to homework 1 Due 9/1/16

BURGESS INEQUALITY IN F p 2. Mei-Chu Chang

Euler s, Fermat s and Wilson s Theorems

Sum and shifted-product subsets of product-sets over finite rings

Florian Luca Instituto de Matemáticas, Universidad Nacional Autonoma de México, C.P , Morelia, Michoacán, México

MULTI-GGS-GROUPS HAVE THE CONGRUENCE SUBGROUP PROPERTY

NONABELIAN GROUPS WITH PERFECT ORDER SUBSETS

Math 324 Summer 2012 Elementary Number Theory Notes on Mathematical Induction

A Collection of MTA ELTE GAC manuscripts

A lattice point problem and additive number theory

arxiv: v3 [math.nt] 29 Mar 2016

Why Bohr got interested in his radius and what it has led to

Sign changes of Fourier coefficients of cusp forms supported on prime power indices

arxiv: v1 [math.co] 13 Sep 2018

APPLICATION OF HOARE S THEOREM TO SYMMETRIES OF RIEMANN SURFACES

On the second smallest prime non-residue

1 2 3 style total. Circle the correct answer; no explanation is required. Each problem in this section counts 5 points.

Tripotents: a class of strongly clean elements in rings

Maximal Class Numbers of CM Number Fields

Homework #2 Solutions Due: September 5, for all n N n 3 = n2 (n + 1) 2 4

A proof of Freiman s Theorem, continued. An analogue of Freiman s Theorem in a bounded torsion group

Expansions of quadratic maps in prime fields

Exponential and character sums with Mersenne numbers

Predictive criteria for the representation of primes by binary quadratic forms

EXPLICIT EVALUATIONS OF SOME WEIL SUMS. 1. Introduction In this article we will explicitly evaluate exponential sums of the form

ON THE ZERO-DIVISOR-CUP-LENGTH OF SPACES OF ORIENTED ISOMETRY CLASSES OF PLANAR POLYGONS. 1. Introduction

ON THE LIMIT POINTS OF THE FRACTIONAL PARTS OF POWERS OF PISOT NUMBERS

Sums of Squares. Bianca Homberg and Minna Liu

Difference Sets are Not Multiplicatively Closed

Marcos J. González Departamento de Matemáticas Puras y Aplicadas, Universidad Simón Bolívar, Sartenejas, Caracas, Venezuela

Products of Factorials Modulo p

arxiv:math/ v1 [math.nt] 20 Mar 2007

Factorization of integer-valued polynomials with square-free denominator

ON INTEGERS EXPRESSIBLE BY SOME SPECIAL LINEAR FORM. 1. Introduction

THE STRUCTURE OF RAINBOW-FREE COLORINGS FOR LINEAR EQUATIONS ON THREE VARIABLES IN Z p. Mario Huicochea CINNMA, Querétaro, México

a mod a 2 mod

Homework #5 Solutions

Sidon sets and C 4 -saturated graphs

Factorization in Integral Domains II

Wilson s Theorem and Fermat s Little Theorem

ON THE RESIDUALITY A FINITE p-group OF HN N-EXTENSIONS

How many units can a commutative ring have?

The minimum rank of matrices and the equivalence class graph

A Note on Indefinite Ternary Quadratic Forms Representing All Odd Integers. Key Words: Quadratic Forms, indefinite ternary quadratic.

ON MATCHINGS IN GROUPS

MULTIPLICATIVE SEMIGROUPS RELATED TO THE 3x +1 PROBLEM. 1. Introduction The 3x +1iteration is given by the function on the integers for x even 3x+1

arxiv:math/ v2 [math.nt] 3 Dec 2003

NOTES ON SIMPLE NUMBER THEORY

SUM-PRODUCT ESTIMATES APPLIED TO WARING S PROBLEM OVER FINITE FIELDS

On non-antipodal binary completely regular codes

Perfect Power Riesel Numbers

CHOMP ON NUMERICAL SEMIGROUPS

Groups of Prime Power Order with Derived Subgroup of Prime Order

THE GROUP OF UNITS OF SOME FINITE LOCAL RINGS I

BOHR CLUSTER POINTS OF SIDON SETS. L. Thomas Ramsey University of Hawaii. July 19, 1994

THE IMPLICIT FUNCTION THEOREM FOR CONTINUOUS FUNCTIONS. Carlos Biasi Carlos Gutierrez Edivaldo L. dos Santos. 1. Introduction

A FUCHSIAN GROUP PROOF OF THE HYPERELLIPTICITY OF RIEMANN SURFACES OF GENUS 2

arxiv: v1 [math.co] 13 May 2016

ON MONIC BINARY QUADRATIC FORMS

Optimal shape and position of the support for the internal exact control of a string

ON THE SEMIPRIMITIVITY OF CYCLIC CODES

ON PERMUTATION POLYNOMIALS OF PRESCRIBED SHAPE

Math 110 HW 3 solutions

VARIETIES OF ABELIAN TOPOLOGICAL GROUPS AND SCATTERED SPACES

32 Divisibility Theory in Integral Domains

arxiv: v2 [math.nt] 29 Jul 2017

Decomposing oriented graphs into transitive tournaments

FIBONACCI DIOPHANTINE TRIPLES. Florian Luca and László Szalay Universidad Nacional Autonoma de México, Mexico and University of West Hungary, Hungary

Finite Field Waring s Problem

Transcription:

AN ADDITIVE PROBLEM IN FINITE FIELDS WITH POWERS OF ELEMENTS OF LARGE MULTIPLICATIVE ORDER JAVIER CILLERUELO AND ANA ZUMALACÁRREGUI Abstract. For a given finite field F q, we study sufficient conditions to guarantee that the set {θ1 x + θ y : 1 x M 1, 1 y M } represents all the nonzero elements of F q. We investigate the same problem for θ1 x θ y and as a consequence we prove that any element in the finite field of q elements has a representation of the form θ x θ y, 1 x, y q 3/4 whenever θ has multiplicative order at least q 3/4. This improves the previous known bound for a question possed by A. Odlyzko. 1. Introduction Let p be a large prime and g a primitive root modulo p. Andrew Odlyzko asked for which values of M the set (1) g x g y (mod p) 1 x, y M, contains every residue class modulo p. He conjectured that one can take M to be as small as p 1/+ɛ, for any fixed ɛ > 0 and p large enough in terms of ɛ. Some results have been obtained in this direction. Rudnik and A. Zaharescu [5], using standard methods of characters sums, proved that one can take M cp 3/4 log p for some c > 0. This range was improved to M cp 3/4 by M. Z. Garaev and K.-L. Kueh [] and independently by S. V. Konyagin [4]. Later, V. C. García [3] showed that c = 5/4 is an admissible constant and the first author [1], using a combinatorial approach, improved the constant to + ε, but for p large enough in terms of ɛ > 0. In this note we exploit properties of Sidon sets, combined with the classic exponential sums techniques, to obtain new results on a generalization of the original problem of Odlyzko. We will no longer study differences of powers of primitive roots in prime fields, but differences of elements of large multiplicative order in arbitrary finite fields F q. We write ord q (θ) for the multiplicative order of θ in F q. Theorem 1. Let θ be an element of F q. If min(ord q (θ), M) q 3/4, then {θ x θ y : 1 x, y M} = F q. 000 Mathematics Subject Classification. 11N69 (11A07 11N5). Key words and phrases. Primitive roots, finite fields, Sidon sets, difference sets. The second author is supported by a FPU grant from Ministerio de Educación, Ciencia y Deporte, Spain. Both authors are supported by grants MTM 011-851 of MICINN and ICMAT Severo Ochoa project SEV-011-0087. 1

J. CILLERUELO AND A. ZUMALACÁRREGUI Applying the previous result when θ is a primitive root we obtain the announced improvment on the problem of Odlyzko. Corollary 1. Let g be a primitive root of F q. If M q 3/4, then {g x g y : 1 x, y M} = F q. One can generalize Theorem 1, by considering the set of sums of powers of two elements in the field. Theorem. Let θ 1, θ be two elements of F q. If then min(ord q (θ 1 ), M 1 / ) min(ord q (θ ), M / ) q 3/, F q {θ x 1 + θ y : 1 x M 1, 1 y M }, F q {θ x 1 θ y : 1 x M 1, 1 y M }. Let us note that, if we consider the case θ 1 = θ and M 1 = M = M the hypothesis in Theorem, say min(ord q (θ), M) q 3/4, are more restrictive that those in Theorem 1. The loss on the constant / in the hypothesis relies on the fact that the set {θ x 1 θ y : 1 x M 1, 1 y M } is no longer symmetric if θ 1 θ or M 1 M. We observe also that 0 may not belong to these sets. If θ 1, θ has order (q 1)/ and q is prime, the elements θ x 1 + θ y are sum of two squares and 0 is not of this form if q 3 (mod 4).. Preliminaries Let G be an abelian group. We recall that a set A G is a Sidon set if all the non zero differences a a, a, a A are distinct. Given a set B, it is usual to denote by r B B (m) to the number of representations of m G in the form m = b b with b, b B. Thus, Sidon sets are those sets A with r A A (m) 1 for all m 0. There are many interesting examples of Sidon sets, but we are interested in those described in the following Lemma. Lemma 1. Let q be a power of a prime and λ a nonzero element of F q. For any given g 1, g, primitive roots of F q, the sets () (3) A (g 1, g, λ) = {(x, y) : g x 1 g y = λ} A + (g 1, g, λ) = {(x, y) : g x 1 + g y = λ} are Sidon sets in G = Z Z. This is a well known fact, see for example [1], but we include the proof for completeness.

AN ADDITIVE PROBLEM IN FINITE FIELDS 3 Proof. We will first show the Sidon condition for A (g 1, g, λ), and the same argument applies to A + (g 1, g, λ). For a fixed m = (m 1, m ) G, m 0, we will show that if there exist a = (x, y), a = (x, y ) A (g 1, g, λ) for which m = a a, then they are uniquely determined by m. Observe that m = a a is equivalent to x = x + m 1 and y = y + m modulo (q 1), which implies that g x 1 g x +m 1 1 (mod q) and g y g y +m (mod q). Combining this observation with the fact that both a and a are in A (g 1, g, λ), we have that (4) g y (g m 1 1 g m ) λ (1 g m 1 1 ). If m 1 0, then y is uniquely determined by equation (4) (and so x, x and y). If m 1 = 0, equation (4) implies that m = 0, which contradicts the assumption m 0. For a real nonzero number x, let us denote by e(x) the complex number e πix. The additive characters ψ in G = Z Z are all of the form ψ r,s (x, y) = e ( ) rx+sy, where 0 r, s q 1 and the character corresponding to r = s = 0 is the principal character. Proposition 1. For any Sidon set A described in Lemma 1 and for any non principal character ψ in G, we have ψ(a) q1/. Proof. We first consider the case () with g 1 = g : A = A (g 1, g 1, λ). Note that for any nontrivial character ψ in G = Z Z we have that (5) ψ(a) = ψ(a a ) = r A A (m)ψ(m) = (r A A (m) 1)ψ(m). a,a A m G m G Since A is a Sidon set, we have that r A A (m) 1 for all m 0 and r A A (0) = A. It follows from (5) that (6) ψ(a) = A 1 ψ(m). m/ A A Thus we need to study the set A A. Observe that the 3(q ) elements of the form (z, 0), (0, z) and (z, z), 1 z q do not belong to A A. Indeed, if (z, 0) = (x + z, y) (x, y) for some (x + z, y), (x, y) A we would have that g x+z g y = g x g y = λ, which is impossible unless z 0 (mod q 1). The same argument applies to the elements of the form (0, z) and (z, z).

4 J. CILLERUELO AND A. ZUMALACÁRREGUI Furthermore, since A A = ( A A + 1) = 3(q ), it follows that those are the only elements m A A. Therefore, for a given ψ = ψ r,s, we have m/ A A q ψ(m) = z=1 ( rz ) q e + q 1 z=1 ( sz ) q e + q 1 z=1 ( (r + s)z ) e 3, q 1 since every such sum is either 1 or q, depending on the values r and s. Combining this bound with the expression in (6), we obtain the desired result ψ(a) ( A + ) 1/ = q 1/. The case () for g 1 g can be reduced the previous one. Given (r, s) (0, 0) let t be the integer such that g t 1 = g. We observe that (x, y) A (g 1, g, λ) (x, ty) A (g 1, g 1, λ). Then, for any a = (x, y) A (g 1, g, λ) we have ψ r,s (x, y) = ψ r,st 1(x, ty). Thus max ψ r,s (a) = max ψ r,st 1(a) q 1/. (g 1,g,λ) (g 1,g 1,λ) The case (3) is easier. It is clear that (x, y) A + (g 1, g, λ) (x, y) + (0, (q 1)/) A (g 1, g, λ) and that ψ(a + (0, (q 1)/)) = ψ(a)ψ(0, (q 1/)). Thus max ψ(a) = + (g 1,g,λ) max ψ(a + (0, (q 1)/)) = (g 1,g 1,λ) max ψ(a) q 1/. (g 1,g 1,λ) As usual, for any set B we define B + B = {b + b : b, b B}. Proposition. Let A be any Sidon set described in Lemma 1 and let B any subset of Z Z. If (B + B) A = then B < q 3/ q + q 1/ + 1/. Proof. The number of pairs (b, b ) B B with b + b A is given by ψ b,b B ψ(b + b a) = B A + 1 ψ ψ 0,0 b,b B ψ(b + b a).

AN ADDITIVE PROBLEM IN FINITE FIELDS 5 Since (B + B) A = it follows from the previous equation and Proposition 1 that B A = 1 ψ( a) ψ(b + b ) ψ ψ 0,0 b,b B 1 ψ( a) ψ(b) b B This implies that B q1/ = q1/ ψ ψ 0,0 ψ ψ 0,0 ψ(b) b B ( B B ). q1/ A + q = (q 1) q 1/ 1/ q + q < 1/ q3/ q + q 1/ + 1/. The easiest way to check the last inequality is multiplying q + q 1/ times q 3/ q + q 1/ + 1/. 3. Proofs of the results We will prove the theorems by a direct application of Proposition to appropriate sets B. 3.1. Proof of Theorem. Let us assume that there exists a fixed nonzero element λ of F q with no solutions to (7) θ x 1 + θ y = λ in F q with 1 x M 1, 1 y M, where (8) min(ord q (θ 1 ), M 1 / ) min(ord q (θ ), M / ) q 3/. Let us denote by n 1 = (q 1)/ord q (θ 1 ) and let g 1 be a generator of F q satisfying θ 1 = g n 1 1. We define n and g analogously. Consider the Sidon set and the set A = A + (g 1, g, λ) B = {(n 1 x, n y) : 1 x M 1 /, 1 y M / }. It is clear that under the previous assumption above we have that (B + B) A =. Then we apply Proposition to this case taking into account that B = min(ord q (θ 1 ), M 1 / ) min(ord q (θ ), M / ) < q 3/ q + q 1/ + 1/ < q 3/ for q, which contradicts (8). The same argument holds for the set of differences by fixing A = A(g 1, g, λ).

6 J. CILLERUELO AND A. ZUMALACÁRREGUI 3.. Proof of Theorem 1. It is clear that the zero element has a representation of the desired form. Let us assume that min(ord q (θ), M) q 3/4 and that there exists a fixed nonzero element λ of F q with no solutions to (9) θ x θ y = λ in F q with 1 x, y M. Let us denote by n = (q 1)/ord q (θ) and let g be a generator of F q satisfying θ = g n. Consider the Sidon set A = A (g, g, λ) and the set B = B 1 B where We claim that B 1 = {(nx, ny) : 1 x, y M/, } B = B 1 + (, ). (10) (B + B) A =. Indeed, any element of B + B is of the form ( ) nx + δ, ny + δ, where δ {0, 1} and 1 x, y M. If one of these elements would belong to A, then nx+δ ny+δ g g = λ. Since g = 1, then either θ x θ y = λ or θ y θ x = λ occur in F q, according to the value of δ. Therefore equation (9) would have a solution. (11) Proposition and (10) imply an upper bound for B : We will get now the lower bound: B < q 3/ q + q 1/ + 1/. (1) B q 3/ q 3/4 + 1/. If M ord q (θ) = > q 3/4, then n { } (nx, ny) : 1 x, y n B1, {( ) } nx +, ny + : 1 x, y n B. Since both sets on the left side are disjoint, we have that q 1 ordq (θ) B n If M < ord q (θ) = n, the sets B 1 and B are disjoint and we have M B =

AN ADDITIVE PROBLEM IN FINITE FIELDS 7 In both cases we have that ( min(ordq (θ), M) q 3/4 B 1 ( = q ) 3/4 1/ ) as we wanted to show. Next we observe that if (11) and (1) hold then = q 3/ q 3/4 + 1/ q < q 3/4 + q 1/. This inequality does not hold for q 16 and it proves the theorem for q in this range. When q < 16, we observe that by assumption min(ord q (θ), M) q 3/4 > q/ (since q/ q 3/4 implies q 64). Suppose that λ D D where D = {θ x : 1 x M} and D = min(ord q (θ), M) > q/. Then D (D + λ) = and we have that which is a contradiction. q D (D + λ) = D = min(ord q (θ), M) > q, Acknowledgments. The authors would like to thank the anonymous referee for his suggestions, that improved the quality of the manuscript. References [1] J. Cilleruelo, Combinatorial problems in finite fields and Sidon sets, Combinatorica, 3 n o 5 (01). [] M. Z. Garaev and Ka-Lam Kueh, Distribution of special sequences modulo a large prime, Int. J. Math. Math. Sci., no.50 (003), 3189 3194. [3] C. V. García, A note on an additive problem with powers of a primitive root., Bol. Soc. Mat. Mexicana (3) 11 n o 1 (005), 1 4. [4] S. V. Konyagin, Bounds of exponential sums over subgroups and Gauss sums, 4th Intern. Conf. Modern Problems of Number Theory and Its Applications, Moscow Lomonosov State Univ., Moscow, 00, 86 114. [5] Z. Rudnik and A. Zaharescu, The distribution of spacings between small powers of a primitive root, Israel J. Math., 10 (000), 71 87. J. Cilleruelo: Instituto de Ciencias Matemáticas (CSIC-UAM-UC3M-UCM) and Departamento de Matemáticas, Universidad Autónoma de Madrid, 8049 Madrid, Spain E-mail address: franciscojavier.cilleruelo@uam.es A. Zumalacárregui: Instituto de Ciencias Matemáticas (CSIC-UAM-UC3M-UCM) and Departamento de Matemáticas, Universidad Autónoma de Madrid, 8049 Madrid, Spain E-mail address: ana.zumalacarregui@uam.es