BEGINNERS' CORNER

Edited by DMITRI THORO

San Jose State College
THE EUCLI DEAN ALGORI THM II
1. I NTRODUCTI ON
In P a r t I [ l ] we s aw t hat t he g r e a t e s t c o mmo n d i v i s o r of . t wo
n u mb e r s coul d be c onve ni e nt l y c o mp u t e d vi a t he f a mous Eu c l i d e a n
a l g o r i t h m. Suppos e t ha t e x a c t l y n s t e ps ( di vi s i ons ) a r e r e q u i r e d t o
c o mp u t e t he g. c. d. of s and t (s > t ) . We t he n ha ve
(1) s = t q , + r , 0 < r < t
(2) t = r , q^ + r , 0 < r
?
< r
(3) r
x
= r
2
q
3
+ x
y
0 < ^ < r
z
(4) r
2
= r
3
q
4
+ r ^ 0 < T^ < r^
(5) r
3
= r
4
q
5
+ r ^
Q < x
5
x
4
( n- 1) r
0
= r ^ q -,+r
1
0 < r , < r
9
N
' n - 3 n - 2 ^ n - 1 n - 1 , n- 1 n - 2
(n) r ~ = r , q + 0 .
x
' n - 2 n - 1 ^n
Si nce e a c h quot i e nt q. > 1, t he a bove e qua t i ons i mp l y
e t c ,
(2
1
) t > r
:
+ r
2
(3' )
r
l - ^
r
2
+ r
3
(4' ) r
2
> r
3 + r 4
(5' )
r
3
> r
4
+ r
5
135
136 F' F. CJNr^Rt ' ^ ' r i i or
w
Apr i l
Fr om (2
1
) and ^ i, - ~, * -' , , ' /, I r om (4
f
)
?
2 r + r >
(2 r . + 2 r J + r_, Si m: ^"Ly, L - -..
t
'i
r
\ ^ -! " r . > (3 r . + 3 r . ) +
2 r . , et c. Continuing In ch.
-1
s JI* u % . .v
j
i<^ * he gener ous abundance
of Fi bonacci number s . Thus
t r , +r > 2r -1-r, > 3r
n
- f 2r , > 5r , +3r
r
1 2 2 j ~~ ..:> 4 *~ 4 5
> , r _ + F
n
r _
"~ n-1 n- 2 n-2 n- 1
2. A BASIC RESULT
Since the r emai nder s form a sti' i^tiy decr eas i ng sequence with
r , the l ast n on- zer o r emai nder .
n- 1 -
c v r
XI-& "* 1 1 - I
Consequently,
n- 1 xi" L n-L ) j - ! '"" n- 1 n-Z n+1
To s ummar i ze, if n di vi si ons a r e r equi r ed to comput e the g. c. d. of
st
s and t, then t i s at l east as l ar ge as the (n + 1) Fi bonacci
number !
3. LAME' S THEOREM
Although the Eucl i dean al gor i t hm i s over 2, 000 year s old, the
following r es ul t was est abl i shed by Gabr i el Lame in 1844.
Theor em
The number of di vi si ons r equi r ed to find the g c. d. of two num-
ber s is never gr eat er than five t i mes the number of di gi t s in the s mal l er
number .
Proof.
Let 4> desi gnat e the golden r at i o. In [2] it was shown t hat
*p
n
= F
n
<t> + F
n
^ | , n=l , 2, 3, . . .
1964 BEGINNERS' CORNER 137
Now si nce 2 > <t> = (1 + ^/5)/Z
9
we see t hat
2F + F , > F 4> + F , or
n i i -1 n n- 1
F > *
n
n+2
Repl aci ng n by n~l and usi ng the "basi c r e s ul t " of the pr ecedi ng
sect i on yi el ds
t > *
n
-
1
.
To compl et e the proof note t hat
(i) if t has d di gi t s then d > log t
(ii) log t > (n-1) log 0
(iii) log 4> > 1/5 .
Thus d > ( n- l ) / 5 or n < 5d .
REFERENCES
1. D. E. Thoro, "The Eucl i dean Al gor i t hm I, " Fi bonacci Quar t er l y,
Vol. 2, No. 1, Febr uar y 1964.
(Note t hat i ne xc e r c i s e s E8 and E10, (F . , , F ) and max (n, F- l )
v
n+1 n n
should be r epl aced by N(F , . , F ) and max N (n, F- l )
^
J
n+1 n'
n
r espect i vel y. )
2. D. E. Thoro, "The Golden Rat i o: Comput at i onal Consi der a-
t i ons, " Fi bonacci Quar t er l y, Vol. 1, No. 3, Oct ober 1963,
pp. 53- 59.
xxxxxxxxxxxxxxxxxxxx