Scribd
Upload
21 views

Advanced Problems.

Maths.

Uploaded by

Charlie Pinedo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
21 views

Advanced Problems.

Maths.

Uploaded by

Charlie Pinedo
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

1981]

381

ADVANCED PROBLEMS AND SOLUTIONS


Edited

by

RAYMOND E. WHITNEY
Lock Haven State

College,

Lock Haven,

PA 17745

Please send all communications


concerning
ADVANCED PROBLEMS AND SOLUTIONS
to RAYMOND E. WHITNEY, Mathematics
Department,
Lock Haven State College,
Lock
Haven, PA 17745.
This department
especially
welcomes problems believed
to be
new or extending
old results.
Proposers
should submit
solutions
or other
information
that will assist
the editor.
To facilitate
their consideration,
the
solutions
should be submitted
on separate
signed sheets
within
two months
after
publication
of the
problems.

PROBLEMS PROPOSED IN THIS ISSUE


H-330

Proposed

by Verner

E. Hoggatt,

Jr.,

San Jose

State

Univ.,

San Jose CA

If 0 is a positive irrational number and 1/6 + 1/8 = 1,


An = [n9], Bn = [n0 3 ], Cn = [n0 2 ] 5
then prove or disprove:
An + Bn + Cn - CBn .
H-331

Proposed

by Andreas

N. Philippou,

American

Univ.

of Beirut,

Lebanon

For each fixed integer k >_ 29 define the k-Fibonacci sequence {/


k)

f<

= 0, f<"

Jn=o by

= 1, and
U> _ [fn-l+

foW

if 2 < <

k,

Letting a = [(1 + /5)/2], show:


(a) /<*> > an'2
(fe)

(b) {/
H-332

)n = 2

Proposed

if n >_ 3;
nas

Schnirelmann density 0,

by David

Zeitlin,

Minneapolis,

MN

Let a = (1 + /5)/2. Let [x] denote the greatest integer function.


that after k iterations (k >_ 1)9 we obtain the identity

Show

[a^ + 2 [ a ^ +2 [a^ + 2 [.]]]] = f(2p+1K2fc+1 ) ^ 2 p + 1 . <P = 0, 1, . . . )


Rem<Vik&: The special case p = 0 appears as line 1 in Theorem 2., p. 309S in the
paper by Hoggatt and Bicknell-Johnson, The Fibonacci
Quarterly
17(4):306-318.
For k = 2 9 the above identity gives
LU

LU

JJ

5(2p + l) /ZP+1

<+(2p+l)

2(2p+l)

382

ADVANCED PROBLEMS AND SOLUTIONS

[Oct.

SOLUTIONS
Con-Vergent
H-308

Proposed by Paul S. Bruckman,


(Vol. 17, No. 4, Dec, 1979)

Corcord,

CA

Let
p
p (a . a . ... 9 a)
[an , a, ..., an] = = 7
r1
2
nJ
qn
qn(al9
a2, ...9 an)
denote the nth convergent of the infinite simple continued fraction
[ax, a 2 , . . . ] , n = 1, 2, ... .
Also, define pQ = 1, q = 0. Further, define
(1)

^n,k = P n (<V a 2 , ..., an)qfc(ax, a 2 , ..., a k )


- p f e ( a 1 5 a 2 , . . . , ak)qn(al9
= p q,

a2, . . . , an)

-p1q90<k<n.

Find a g e n e r a l formula f o r Wntk.


Solution

by the

proposer.

R e c a l l t h a t t h e p ' s and q n f s s a t i s f y t h e b a s i c
(2)

rn

+1

= an + 1rn + rn_19

recursion

n 1, 2

Also-, the following relations are either obvious or well known:


(3)

Wn9n = 0;

(*)

= (-1)"'

tf,n-i

(5)

tfn,n.2

> !;

= (-D " an, n > 2.

[See Niven and Zuckerman* An Introduction


to the Theory of Numbers9
(New York: Wiley, 1972), Theorem 7.5, for a proof of (4) and (5).]
We show, by strong induction, that

3rd ed.

<6>
".k = <-1>k+1P-k-l K +2>fl*+ 3' -. ">
Let 5 denote the set of positive integers n such that (6) holds for 0 <_ k < n.
Setting n = 1 in (4) yields W 0 = -1 = (-l)0+1p0; hence, 1 e S. Suppose that
for some integer m >_ 2, 1, 2, .. ., 772 e S. By (4) and (5), we have:

Wm+1,m = (-l)m+1 = (-l)m+1 p 0 , and Wm + i,m-i

(7)

= (-D\tl. r

Also, if 0 < fem - 2,


W'm + l . k - P + l < ? k

- Pfc<? m

<am + lP

+ Pm-l><7k ~ P k ( a + l<7m + ? - l >

= m + i(P B ? fc " P k ^ ) + P r a A " P , ^ . , = am + 1m,

+ <V1(k

[ u s i n g (1) and ( 2 ) ] . Hence, by t h e i n d u c t i v e h y p o t h e s i s and ( 2 ) ,


&m + l , f c

=:

C" 1 )

m+lPw~-fe-l(a7< + 2 9

^ ^ Vm -*7< ^ .
r

Thus, u s i n g (7) and ( 8 ) ,

-. m + i)-

rc)

(_1)

Pm-fc-2( a fe + 2

m-l)

1981]

ADVANCED PROBLEMS AND SOLUTIONS

<9>

^ + i . k " (- 1 )* +1 P- k K +2

383

am+1),0<fe<m,

which is equivalent to the statement (jn + 1) e S.

Hence9

1, 2, . . . , w i 5 ^ ( m + 1 ) e 5 .
By induction9 (6) is proved.
Fibonacci and Lucas Are the Greatest Integers
H-310

Proposed by Verner E. Hoggatt,


(Vol. 17, No. 4, Dec, 1979)

Jr.,

San Jose

State

Let a = (1 + v/5)/29 [na] = an9 and [na2] = bn.


(a)

Show that if n = F2m+1,

(b)

Show that if n = F 2m,

(c)

Show that if n = L 2 m , then a n = L2m + i

(d)

Show that if n = 2 m + 1 , then a n = L2m

Solution

by Paul S. Bruckman,

then a n = F2m

then a n = F2m

+1

+2

Univ.,

San Jose, CA

Clearly an + n = bn.
and n = F2m + 3.

- 1 and bn = F 2m +2 - 1.
and

+2

^n = L2m+2-

- 1 and bn = L 2 m + 3 - 1.

Corcord, CA

We begin by noting that

- O F = 4={a n+1 - 3 n + 1 - a(an - 3n)}

n +l

/5

= 4=(a n + 1 - 3 n + 1 - an + 1 - 3 n - x )
= -3n//5(3 - a ) ,
or
3n=^

(1)
v

+1

A l s o , aln - Ln+l
(2)

, - aF.

rc+1

n+1

= a ( a " + 3") - ( a " + 3


eV5=a -

) = -P B (B - a ) , o r
Ln+l.

Since - 1 < 3 < 0 , t h u s 0 < 3 2 n 1 and - 1 < 3 2 n + 1 < 0 (n _> 0 ) . Hence, u s i n g (1)
0

< F2n+1

aF

2n

a n d

~l

< F 2 n + 2 " *F2n+l

< 5

note that equality is attained above if and only if n = 0.


F2n

+1

- l i F 2 n < F2n + 1

and F2n + 2 < aF2n+1

< F 2n

+2

Therefore,

+ 1 (n > 0 ) .

It follows that
(3)

[aF2n]

-F2n+1

- 1, and

(4)

[oP2+ i] = 2n + 2 < >.0).


Now (3) i m p l i e s [ a 2 F 2 ] = [ ( 1 + a ) F 2 ] = F'2n + [aF2n]
2

(5)

[ F2J
2

A l s o , [a F2n+1]
(6)

- F 2 n + 2 - 1.

- F 2 n + 1 + [aF2n+1]
2

= F2n + F2n + l~ 1, o r

= F2n+1

+ F2n+2,

or

[a ^2n + J = 2n + 3Note that (4) and (6) are equivalent to (a) of the original problem; also,
(3) and (5) are equivalent to (b) of the original problem.
In order to prove (c) and (d)s we proceed similarly, using the result in
(2). We need only observe that |3n/5*| < 1 for n >_2. The desired results then

384

ADVANCED PROBLEMS AND SOLUTIONS

Oct. 1981

follow, as before, for all values of n except for possibly n - 0; however, a


quick inspection shows that the results also hold for n = 0, i.e.,
(7)

[aL2n]

= L 2 n + 1 , [aL 2n+1 ] = L2n+2

- 1,

which imply the other two results.


Commtnt by Bob Prielipp,

University

of Wisconsin-Oshkosh,

WI

Sharp-eyed readers will find that this problem can be solved easily by using the following four lemmas established in the article "Representations of
Integers in Terms of Greatest Integer Functions and the Golden Section Ratio"
by Hoggatt and Bicknell-Johnson [The Fibonacci
Quarterly
17(4):306-318].
Lemma 1 (p. 308):

[dFn] = Fn+1>
[aFn]

Lemma 2 (p. 308) :

= Fn + 2 - 1, n even, n j> 2.

[aLn] = Ln + 1 for n even, if n _> 2;


[aLn]

Lefflma 7 (p. 375) :

= Fn + 1 - 1, w even, n >. 2.

[a2Fn ] = Fn + 2, n odd, n >_ 2;


[a2Fn]

Lemma 6 (p. 315) :

n odd, n >_ 2;

= n+1 - 1 for n odd, if n >_ 3.

[a Ln] = Ln + 2 if n is even and n _> 2;


[a2Ln] = n+2 - 1 ifftis odd and n _> 1.

Also solved

by Bob Prielipp,

G, Wulczyn,

and the

proposers.

CORRECTIONS
1.

The problem solved in Vol. 18, No. 2, April 1980 is H-284 not H-285.

2.

H-315 as it appeared in Vol. 18, No. 2, April 1980 had several misprints in
it. A corrected version is given below.

H-315

Proposed by D. P. Laurie,
Sciences,
Pretoria,
South

National
Africa

Research

Institute

for

Mathematical

Let the polynomial P be given by


P O O = zn + an_1zn~1

+ an_2zn~2

+ ... + az + aQ

and let zls z2, ..., zn be distinct complex numbers. The following iteration
scheme for factorizing P has been suggested by Kerner [1]:
,

Il(Zi

u --

J-J

^-J

~ 3j)

J-l

Prove that if _.zi


J=I

***
- ^n_1 then a l s o 2_, ^i
i- 1

"an-l-

Reference
1.

I. Kerner. "Ein Gesamtschrittverfahren zur Berechnung der Nullstellen von


Polynomen." Numer. Math. 8 (1966):290-94.

You might also like