• 検索結果がありません。

SHORT PROOFS OF THEOREMS OF LEKKERKERKER AND BALLIEU

N/A
N/A
Protected

Academic year: 2022

シェア "SHORT PROOFS OF THEOREMS OF LEKKERKERKER AND BALLIEU"

Copied!
4
0
0

読み込み中.... (全文を見る)

全文

(1)

nternat. J. Math. & Math. Sci.

Vol. 5 No.

3

(1982)609-612

609

SHORT PROOFS OF THEOREMS OF LEKKERKERKER AND BALLIEU

MAX RIEDERLE

Eberhardstr. 14 79 Ulm/Donau West Germany

(Received October 16, 1981)

ABSTRACT. For any irrational number let A() denote the set of all accumulation points of

{z:

z=q(q- p), p

,

q

- {0},

p and q relatively

prime}.

In this

paper the following theorem of Lekkerkerker is proved in a short and elementary way:

The set A() is discrete and does not contain zero if and only if is a quadratic irrational. The procedure used for this proof simultaneously takes care of a theorem of Ballieu.

KEY WORDS AND PHRASES. Lekkerkerker s Theorem, Approximation of

numbeJ

Quadratic I

atio

na

1980

MATHEMATICS SUBJECT CLASSIFICATION CODES. 10F05, 10F35.

1. INTRODUCTION.

This paper is easily readable by anyone familiar with the elements of continued fractions, as far as

Lagrange’s

theorem on the periodic representation of quadratic irrationals. Throughout this paper, denotes an irrational number which is repre- sented by the regular continued fraction

[b0,bl,b

2

[bo,b

I

,bn_l,n]

and A() stands for the set of all the real accumulation points of {z: z q(q-p), p e q

- {0},

p and q relatively

prime}.

Obviously,

A()

describes those

Dirichlet approximation qualities which occur infinitely often. Furthermore, for any sequence (a) let H(a denote the set of all its limit points and for x e R

n n

and g > 0 set B(x,g) (x g, x

+

g). The main purpose of this paper is to gJ’e a simple proof of the following theorem of Lekkerkerker

[i]

(cf. also [2])"

(2)

610 M. RIEDERLE

The set

A()

is discrete and does not contain zero if and only if is a quadratic irrational.

The proof of the sufficient part of the theorem mainly depends on the irre- ducible polynomial of

,

whereas the necessary part is a consequence of the relation between

A()

and the sequence

(n)

and simultaneously establishes the following

theorem of Ballieu

[3]:

The set

H(

is finite and () is bounded if and only f is a quadratic

n n

irrational.

Finally for any quadratic irrational we shall show how to evaluate A() in an easy way.

2. BASIC FORMULAS.

In this section we state the formulas used in the sequel. Let A /B denote n n the n-th convergent of

[bo,bl,b

2,... where

An

and

Bn

are relatively prime. Set On

Bn/Bn_

1 and put

6n Bn(Bn- An)"

Then the following formulas hold for all n

e]O:

6

(-l)n

n

n+l + i/0n

(2.1)

(-)

n-i

6n_

I

On + I/n+

I

(2.2)

i+i+46

6

n n-i n

n+l 26

n (-I) (2.3)

PROOF. Equation (2.1) is an easy consequence of the well known identity

- An/B

n

(-l)n/(Bn[Bnn+l_+ Bn_,l)),formula

(2.2) can be derived from (2.1) when

b

+

i/ and, finally (2 3) can using the identities

n

bn

+ i/n+ I

and On n

0n-i

be obtained when combining (2.1) and (2.2).

3. PROOF OF

LKKKERKERKER’S

THEOREM.

(1) Suppose is a quadratic irrational. There exists an indefinite quadratic form f(x,y) ax2

+

bxy

+

cy2

with a,b,c e

.

and

f(,l)

O. If denotes the algebraic conjugate of then if follows by Vieta’s theorem that

f(p,q) a(p q)(p q) aq(q- p)( p/q) (3.1)

for all (p,q) (

.

x(.-

{0})

which implies that

(3)

THEOREMS OF LEKKERKERKER AND BALLIEU 611

q(q- p) f(P,q)

a(- p/q) (3.2)

When using the notation

{(x,y):

x,y

.,

y

#

0, x and y relatively

prime}

equation (3.2) implies that

A() (3 3)

a(- E)

since if

(Pn’qn)

Z with

nlmoo qn(qn

p)

A()

then

nlmoo Pn/qn 6.

Clearly,

f(Z) and we conclude that A($) is discrete. Now from (3.1) we can see that f(p,q)

#

0 for all (p,q) e and hence it follows by (3.3) that 0 A().

(ii) Suppose that A() is discrete and 0 A(). From equation (2.1) we can see that

16nI -<

1 for all n

q0"

Therefore and since all the numbers

n

are

distinct,

H(n)

is a compact subset of

A()

and hence

H(n

is finite and 0

H(n).

Now by (2.3) it is easy to see that

H( n)

is finite and

(n)

is bounded. Therefore,

in order to complete the proof, it suffices to prove Ballieu’s theorem.

4. PROOF OF

BALLIEU’S

THEOREM.

(i) Suppose that

($n,)

is bounded and

H(n)

is finite say

H( n) {z

I zm It follows from the identity

n [bn’bn+l’’’"

that there exists a k lq such that

bn

< k for all n IN, and hence

bn +

i/k <

n

<

bn +

i i/(k

+

i). Therefore, the

set

H(n)

n -is empty and we can find a number g > 0 such that

th

sets

B(zv,g)

are pairwise disjoint and contained in]R-Z.

Let l(z) denote the greatest integer not exceeding z and for z Zput

I(z) (z l(z))-i Clearly,

n+l I(n

for all n

IN0"

Also the function I is continuous on

H(n)

therefore

I(H(n) H(n)

and we can find a

6,

0 < 6 < g,

such that

I(B(z,$)) = B(I(z),g)

for all

v

{i

m}

There exists a number nO

m

such that

n

o

B(Zl,)

and

n

=I

B(z,)

for all n > no Therefore, when

writing

(P)

for the p-th composition map of

,

we obtain by induction that

n +

p e

B(I(P)(zl)’6)

for all p

e0"

Since

(H(n)) H(n)

and

H(n

is finite,

O

the sequence

((P)(zl)),

p

lq0,

is periodic. From the identities

bn l(n + p)

o o

I(T(p)

p)

(Zl))

we conclude that the sequence (bn

+

P

0’

is periodic and thus,

O

by

Lagrange’s

theorem, is a quadratic irrational.

(ii) The other direction of Ballieu’s theorem is an easy consequence of

(4)

612 i. RIEDERLE

Lagrange’s

theorem.

5. CONCLUDING REkRKS.

The inclusion in (3.3) is actually an equality. In order to prove this, we need the following well known theorem (cf. [4], p. 22-23):

Let f(x,y) be an indefinite quadratic form with integer coefficients and let be one of its roots. Then for any pair (p,q) e Z

x(-{0})

there are infinitely many relatively prime integers

Pn’ qn

such that

f(pn,qn

f(p,q) for all n

and

nlm (qn

p) O.

In fact, this result combined with (5) and (6), leads to the following:

THEOREM. Suppose that is a quadratic irrational, say

f(,l)

0 for some indefinite quadratic form f(x,y) with integer coefficients. Moreover, let

E

be the algebraic conjugate of

.

Then

A() f(z)

f(l,0)(E- )

REFERENCES

i. LEKKERKERKER, C.G. Una questione di approssimazione diofantea e una proprieta caratteristica dei numeri quadratici I, II. Atti Accad. Naz. Lincei.

Rend. CI. Sci. Fis. [at. Nat. 21 (1956) 179-185, 257-262.

2. JURKAT, W.B. and PERERIMHOFF, A. Characteristic approximat’ion properties of quadratic irrationals, Intern. J. of Math. and Math. Sci. i (1978) 479-496.

3. BALLIEU, R. Sur des suites de nombres

lies

$ une fraction continue

rgulire,

Acad. Roy. Belg. Bull. CI. Sci. 29 (1943) 165-174.

4. CASSELS, J.W.S. An Introduction to Diophantine Approximation. Cambridge Univ.

Press, Cambridge, 1965.

5. PERRON, O. Die Lehre von den

Kettenbrchen.

Teubner Verlag, Stuttgart, 1954.

参照

関連したドキュメント

lattice points, ellipsoids, rational and irrational quadratic forms, pos- itive and indefinite quadratic forms, distribution of values of quadratic forms, Oppenheim

In this paper the algebraic approach to autonomous homogeneous quadratic continuous systems of the form x Qx and autonomous homogeneous quadratic discrete dynamical systems of

Using the fixed point method, we prove the generalized Hyers-Ulam stability of the following additive-quadratic-cubic-quartic functional equation f x 2y f x − 2y 4f x y 4f x − y

An important problem in the theory of quadratic forms is to determine when an anisotropic quadratic form ' over F becomes isotropic over the function eld F ( ) of another form.

SIMSEK, q-Bernoulli numbers and polynomials associated with multiple q-zeta functions and basic L-series,

In [11] the authors classify all centers of the family of the BiLi´ enard systems of degree five and find the maximum number of limit cycles which can bifurcate from a fine focus

Nagaich, “Constancy of holomorphic sectional curvature in indefinite almost Hermitian manifolds,” Kodai Mathematical Journal, vol. Hervella, “Curvature of indefinite almost

We study the Euler equation for Neumann and Dirichlet problems asso- ciated to nonconvex functionals defined on the space of functions with bounded variation and satisfying a safe