REDUCTION MODULO p. We will prove the reduction modulo p theorem in the general form as given by exercise 4.12, p. 143, of [1].

Similar documents
THERE ARE NO POINTS OF ORDER 11 ON ELLIPTIC CURVES OVER Q.

Math 261 Exercise sheet 2

(2mn, m 2 n 2, m 2 + n 2 )

where a is any ideal of R. Lemma Let R be a ring. Then X = Spec R is a topological space. Moreover the open sets

Polynomials. 1 More properties of polynomials

FACTORIZATION IN KRULL MONOIDS WITH INFINITE CLASS GROUP

Foundations of Arithmetic

LECTURE V. 1. More on the Chinese Remainder Theorem We begin by recalling this theorem, proven in the preceeding lecture.

where a is any ideal of R. Lemma 5.4. Let R be a ring. Then X = Spec R is a topological space Moreover the open sets

ALGEBRA HW 7 CLAY SHONKWILER

Introduction to Arithmetic Geometry Fall 2013 Lecture #25 12/05/2013

a b a In case b 0, a being divisible by b is the same as to say that

APPENDIX A Some Linear Algebra

FINITELY-GENERATED MODULES OVER A PRINCIPAL IDEAL DOMAIN

Linear, affine, and convex sets and hulls In the sequel, unless otherwise specified, X will denote a real vector space.

28 Finitely Generated Abelian Groups

Math 101 Fall 2013 Homework #7 Due Friday, November 15, 2013

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix

Smarandache-Zero Divisors in Group Rings

The Ramanujan-Nagell Theorem: Understanding the Proof By Spencer De Chenne

Problem Solving in Math (Math 43900) Fall 2013

1 Matrix representations of canonical matrices

ALGEBRA MID-TERM. 1 Suppose I is a principal ideal of the integral domain R. Prove that the R-module I R I has no non-zero torsion elements.

THE CHINESE REMAINDER THEOREM. We should thank the Chinese for their wonderful remainder theorem. Glenn Stevens

SL n (F ) Equals its Own Derived Group

REAL ANALYSIS I HOMEWORK 1

MATH Sensitivity of Eigenvalue Problems

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0

18.781: Solution to Practice Questions for Final Exam

MTH 819 Algebra I S13. Homework 1/ Solutions. 1 if p n b and p n+1 b 0 otherwise ) = 0 if p q or n m. W i = rw i

Example: (13320, 22140) =? Solution #1: The divisors of are 1, 2, 3, 4, 5, 6, 9, 10, 12, 15, 18, 20, 27, 30, 36, 41,

Section 8.3 Polar Form of Complex Numbers

Affine transformations and convexity

12 MATH 101A: ALGEBRA I, PART C: MULTILINEAR ALGEBRA. 4. Tensor product

COMPLEX NUMBERS AND QUADRATIC EQUATIONS

A p-adic PERRON-FROBENIUS THEOREM

2.3 Nilpotent endomorphisms

Polynomials. 1 What is a polynomial? John Stalker

Graph Reconstruction by Permutations

Math 594. Solutions 1

DISCRIMINANTS AND RAMIFIED PRIMES. 1. Introduction A prime number p is said to be ramified in a number field K if the prime ideal factorization

Lecture 5 Decoding Binary BCH Codes

Problem Do any of the following determine homomorphisms from GL n (C) to GL n (C)?

P.P. PROPERTIES OF GROUP RINGS. Libo Zan and Jianlong Chen

On the Reducibility of Cyclotomic Polynomials over Finite Fields

Formulas for the Determinant

5 The Rational Canonical Form

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal

( 1) i [ d i ]. The claim is that this defines a chain complex. The signs have been inserted into the definition to make this work out.

First day August 1, Problems and Solutions

Semilattices of Rectangular Bands and Groups of Order Two.

MATH 241B FUNCTIONAL ANALYSIS - NOTES EXAMPLES OF C ALGEBRAS

International Journal of Algebra, Vol. 8, 2014, no. 5, HIKARI Ltd,

8.6 The Complex Number System

PHYS 705: Classical Mechanics. Calculus of Variations II

A CHARACTERIZATION OF ADDITIVE DERIVATIONS ON VON NEUMANN ALGEBRAS

On the partial orthogonality of faithful characters. Gregory M. Constantine 1,2

Representation theory and quantum mechanics tutorial Representation theory and quantum conservation laws

arxiv: v1 [math.co] 1 Mar 2014

INTERVAL SEMIGROUPS. W. B. Vasantha Kandasamy Florentin Smarandache

Fixed points of IA-endomorphisms of a free metabelian Lie algebra

Christian Aebi Collège Calvin, Geneva, Switzerland

Character Degrees of Extensions of PSL 2 (q) and SL 2 (q)

Kernel Methods and SVMs Extension

Kuroda s class number relation

Generalized Linear Methods

Complex Numbers. x = B B 2 4AC 2A. or x = x = 2 ± 4 4 (1) (5) 2 (1)

On intransitive graph-restrictive permutation groups

NOTES ON SIMPLIFICATION OF MATRICES

NUMERICAL DIFFERENTIATION

p-adic Galois representations of G E with Char(E) = p > 0 and the ring R

DIFFERENTIAL FORMS BRIAN OSSERMAN

THE CLASS NUMBER THEOREM

Solutions for Tutorial 1

inv lve a journal of mathematics 2008 Vol. 1, No. 1 Divisibility of class numbers of imaginary quadratic function fields

Another converse of Jensen s inequality

More metrics on cartesian products

Canonical transformations

= s j Ui U j. i, j, then s F(U) with s Ui F(U) G(U) F(V ) G(V )

2 More examples with details

A combinatorial problem associated with nonograms

One-sided finite-difference approximations suitable for use with Richardson extrapolation

NOTES FOR QUANTUM GROUPS, CRYSTAL BASES AND REALIZATION OF ŝl(n)-modules

SUCCESSIVE MINIMA AND LATTICE POINTS (AFTER HENK, GILLET AND SOULÉ) M(B) := # ( B Z N)

arxiv: v6 [math.nt] 23 Aug 2016

HOPF ALGEBRAS WITH TRACE AND CLEBSCH-GORDAN COEFFICIENTS. 1. Recollections and the problem

Modelli Clamfim Equazione del Calore Lezione ottobre 2014

The internal structure of natural numbers and one method for the definition of large prime numbers

Lecture 10 Support Vector Machines II

Restricted Lie Algebras. Jared Warner

Problem Set 9 Solutions

CALCULUS CLASSROOM CAPSULES

Section 3.6 Complex Zeros

IRREDUCIBILITY OF ELLIPTIC CURVES AND INTERSECTION WITH LINES.

New York Journal of Mathematics. Characterization of matrix types of ultramatricial algebras

On the Nilpotent Length of Polycyclic Groups

INTERSECTION THEORY CLASS 13

= = = (a) Use the MATLAB command rref to solve the system. (b) Let A be the coefficient matrix and B be the right-hand side of the system.

Perfect Competition and the Nash Bargaining Solution

Distribution det x s on p-adic matrices

Transcription:

REDUCTION MODULO p. IAN KIMING We wll prove the reducton modulo p theorem n the general form as gven by exercse 4.12, p. 143, of [1]. We consder an ellptc curve E defned over Q and gven by a Weerstraß equaton: E : y 2 = x 3 + ax 2 + bx + c where a, b, c Z. To E we have attached the dscrmnant: D = 4a 3 c + a 2 b 2 + 18abc 4b 3 27c 2. Fx a prme number p, and recall that to p we have attached the p-adc valuaton ord p on Q as well as the subrng R p := fx Q j ord p (x) 0g. For x R we denote by x the reducton modulo p of x; thus, x s an element of F p. Defnton 1. The reducton modulo p of E s the cubc curve Ē defned over F p by the equaton: y 2 = x 3 + āx 2 + bx + c. When s Ē an ellptc curve over F p? By our crtera, ths s so precsely when:.e., f and only f: p 2 and D = 4ā 3 c + ā 2 b 2 + 18ā b c 4 b 3 27 c 2 0 p 2 and p D = 4a 3 c + a 2 b 2 + 18abc 4b 3 27c 2. So, there are only fntely many bad prmes p for whch the curve Ē s sngular. For p 2D we wll consder the reducton modulo p map E(Q) Ē(F p ) denoted by P P as follows: If P E(Q) we can wrte P = (x, y, z) wth x, y, z R p and at least 1 a unt n R p ; the vector (x, y, z) s then unquely determned up to multplcaton by a unt n R p, and hence the reducton: P = ( x, ȳ, z) s well-defned as a pont n P 2 (F p ). Clearly, P Ē(F p ). (Alternatvely, wrte P = (x, y, z) wth x, y, z Z and gcd(x, y, z) = 1. Wth ths requrement the vector (x, y, z) s unquely determned up to sgn). 1

2 IAN KIMING Recall our prevous notaton: For ν N: E(p ν ) = f(x, y) j ord p (x) 2ν, ord p (y) 3νg f0g = f(x, y) j ord p (x) 2νg f0g = f(x, y) j ord p (y) 3νg f0g, and that ord p (x) (resp. ord p (y)) must have shape 2ν 0 (resp. 3ν 0 ) f (x, y) E(Q) wth ord p (x) < 0 (resp., wth ord p (y) < 0). Theorem. Assume that p s such that the reduced curve Ē s an ellptc curve over F p,.e., assume that p 2D. The reducton modulo p map s a homomorphsm: wth kernel: E(Q) Ē(F p ) E(p) = f(x, y) j ord p (x) < 0g f0g = f(x, y) j ord p (y) < 0g f0g. Proof. Clearly 0 = 0 Ē(F p ). Let P = (x, y, 1) E(Q). If x R p then y R p, and so P = ( x, ȳ, 1) 0. On the other hand, f ord p (x) < 0 then ord p (x) and ord p (y) have shape: ord p (x) = 2ν, ord p (y) = 3ν for some ν N. But then P = (p 3ν x, p 3ν y, p 3ν ); snce p 3ν x, p 3ν R p p, but p 3ν y s a unt n R p we then deduce P = (0, 1, 0) = 0 Ē(F p ). So we certanly have for P E(Q) that: P = 0 P E(p). We now only have to show that the reducton modulo p map s a homomorphsm. To show that the reducton modulo p map s a homomorphsm we must show that f P 1, P 2, P 3 E(Q) wth P 1 + P 2 + P 3 = 0 then P 1 + P 2 + P 3 = 0. Assume then that P 1, P 2, P 3 E(Q) wth P 1 + P 2 + P 3 = 0. We easly see that ( P ) = P for P E(Q). From ths we see that we may well assume P 0 for = 1, 2, 3. Wrte then P = (x, y ) for = 1, 2, 3. 1. Suppose frst that 2 of the ponts P s n E(p), say P 1, P 2 E(p). Snce E(p) as prevously shown s a subgroup of E(Q) we can then conclude that P 3 = (P 1 + P 2 ) E(p). But then P = 0 for = 1, 2, 3, and thus trvally P 1 + P 2 + P 3 = 0. 2. Suppose then that exactly 1 of the ponts P s n E(p), say P 3 E(p). 2 a. If addtonally P 1 P 2 then the x-coordnate x 3 of the pont P 3 = (P 1 + P 2 ) s: ( ) 2 y2 y 1 x 3 = a x 1 x 2. x 2 x 1

REDUCTION MODULO p. 3 As ord p (x 3 ) < 0 and as now x 1, x 2, y 1, y 2 R p we must then necessarly have ord p (x 2 x 1 ) > 0,.e.: Now consder the dentty: x 1 x 2 (p). (y 2 + y 1 )(y 2 y 1 ) = y 2 2 y 2 1 = x 3 2 x 3 1 + a(x 2 2 x 2 1) + b(x 2 x 1 ) ; snce the rght hand sde has shape (x 2 x 1 ) A wth A R p, we can conclude that f we had y 2 y 1 (p) then we could deduce the mplcaton: ord p (x 2 x 1 ) s ord p (y 2 y 1 ) s for any s N; but ths would mply ord p ( y2 y 1 x 2 x 1 ) 0 and hence the contradcton ord p (x 3 ) 0. So n addton to the congruence x 1 x 2 (p) we have also y 2 y 1 (p). But then: P 2 = ( x 2, ȳ 2 ) = ( x 1, ȳ 1 ) = P 1, and P 1 + P 2 + P 3 = P 1 + P 2 = 0 as requred. 2 b. Suppose now nstead that P 1 = P 2. Snce P 3 0 the x-coordnate x 3 of P 3 = 2P 1 s now gven by: ( f ) 2 (x 1 ) x 3 = 2x 1 a, 2y 1 where f(x) := x 3 + ax 2 + bx + c. We conclude that ord p ( f (x 1) 2y 1 ) ord p (y 1 ) = ord p (2y 1 ) > 0 (here we used that p 2). But then P 1 = ( x 1, ȳ 1 ) = ( x 1, 0) s a pont of order 2; we conclude that P 1 + P 2 + P 3 = 2 P 1 = 0. < 0, hence 3. Suppose now fnally that P E(p) for = 1, 2, 3. So, the coordnates x, y are all n R p. Snce P 1 +P 2 +P 3 = 0 the ponts P 1, P 2, P 3 are precsely the ponts of ntersecton (counted wth multplctes) between E and some lne L. Let: αx + βy + γz = 0 be an equaton for L. We may assume that α, β, γ are n R p and at least 1 of them a unt n R p, say γ s a unt (as wll be seen, the other cases are analogous to deal wth). We can then assume an equaton for L of form: wth λ, ν R p. z = λx + νy Now let F (x, y, z) = 0 be a homogeneous equaton of degree 3 and coeffcents n R p that defnes E. E.g.: F (x, y, z) := x 3 + ax 2 z + bxz 2 + cz 3 y 2 z.

4 IAN KIMING As the ponts P = (x, y ), = 1, 2, 3, are the ponts of ntersecton between E and L, we must have: ( ) F (x, y, λx + νy) = ξ (xy 1 yx 1 )(xy 2 yx 2 )(xy 3 yx 3 ) for some non-zero, ratonal ξ. Suppose that we had ord p (ξ) < 0. Puttng then s := ord p (ξ) we would have s > 0 and the number p s ξ would be a unt n R p ; snce F has coeffcents n R p we could deduce that the polynomal: p s ξ (xy yx ) = p s F (x, y, λx + νy) has all coeffcents n R p ; but then: p s ξ (xȳ y x ) = 0 whence (snce p s ξ were a unt n R p ) (xȳ y x ) = 0 for at least 1 ; but ths means x = ȳ = 0 for that whch s mpossble, as 1 = λx + νy. So we have ξ R p. If we had ξ = 0 we would deduce from ( ) that: F (x, y, λx + νy) = 0 where F denotes the (coeffcentwse) reducton of F ; but that would mean that Ē would contan the whole lne z = λx + νy, somethng that we know s not possble. So we have n fact that ξ s a unt n R p. Then: F (x, y, λx + νy) = ξ (xȳ y x ) wth ξ 0. Snce clearly F = 0 s an equaton defnng Ē over F p, we can now conclude that the ponts P = ( x, ȳ ) are the ponts (counted wth multplctes) of ntersecton between Ē and the lne gven over F p by the equaton z = λx + νy. We conclude that P 1 + P 2 + P 3 = 0. Corollary 1. For p a prme such that p 2D the reducton modulo p map gves an njectve homomorphsm: E(Q) tors Ē(F p ). Proof. The statement follows from the theorem snce the subgroup E(p) s torson free. Corollary 2. If p s a prme such that p 2D then #E(Q) tors j #Ē(F p ). Proof. E(Q) tors s somorphc to a subgroup of the fnte abelan group Ē(F p ). By the last corollary, f we have prmes p 1,..., p t, none of whch dvdes 2D, then #E(Q) tors j gcd(#ē(f p1 ),..., #Ē(F pt )).

REDUCTION MODULO p. 5 In practce, ths often gves a very effcent method of determnng #E(Q) tors and n fact E(Q) tors tself. Let us look at a couple of examples. If p 2D one can compute Ē(F p ) wth the program PARI. The relevant commands are: gp : Start PARI E = ellnt([a 1, a 2, a 3, a 4, a 5 ]) : Intalze the curve E : y 2 + a 1 xy + a 3 y = x 3 + a 2 x 2 + a 4 x + a 6. ellap(e, p) : Compute the number a p := p + 1 #Ē(F p ) (for p 2D). Example 1: If E s the curve y 2 = x 3 + 80 then D = 2 8 3 3 5 2. One computes a 7 = 5, a 11 = 0, so that #Ē(F 7 ) = 8 ( 5) = 13 and #Ē(F 11 ) = 12 0 = 12. Snce gcd(12, 13) = 1 we deduce that E(Q) tors = 0. Example 2: Consder the curve E gven by y 2 = x 3 + x. D = 4. E has the obvous torson pont (0, 0) of order 2. The dscrmnant s One computes: #Ē(F 3 ) = 4, #Ē(F 5 ) = 4, #Ē(F 7 ) = 8,.... In fact, one can show that #Ē(F p ) s always dvsble by 4 whenever p s an odd prme. But we clam that E(Q) tors = f0, (0, 0)g and that ths can stll be deduced from the above theorems: For let us for p = 3, 5 look not only at the order of the group #Ē(F p ) but also at ts structure: We fnd: and hence that Ē(F 3 ) = Z/4. Also: and consequently Ē(F 5 ) = Z/2 Z/2. Ē(F 3 ) = f0, (0, 0), (2, 1), (2, 2)g Ē(F 5 ) = f0, (0, 0), (2, 0), (3, 0)g Snce E(Q) tors embeds nto both Ē(F 3 ) and Ē(F 5 ) we deduce that the fnte abelan group E(Q) tors s both cyclc and has exponent a dvsor of 2. As E(Q) tors contans n any case a subgroup of order 2, namely the subgroup generated by (0, 0), the clam follows. References [1] J. H. Slverman, J. Tate: Ratonal ponts on ellptc curves. Undergraduate Texts n Mathematcs, Sprnger-Verlag 1994. Department of Mathematcs, Unversty of Copenhagen, Unverstetsparken 5, DK- 2100 Copenhagen Ø, Denmark. E-mal address: kmng@math.ku.dk