Consider all pairs of positive integers ( G , L ) such that 2 ≤ G ≤ L ≤ 1 0 0 0 . For each pair, let there be N G , L ordered pairs of positive integers ( a , b ) such that G = g cd ( a , b ) and L = l c m ( a , b ) .
What is the largest possible value of N G , L ?
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
Nicely done!
thanks!
Since G is the g cd , we can uniquely write: a = G A b = G B for some coprime integers A , B . Then L = G A B , from the definitions of g cd and lcm .
So we need to find the maximum number of ways that a value A B can be broken up into a pair of coprime factors, subject to the constraint that G A B ≤ 1 0 0 0 .
Note that G doesn't have to be coprime to A , B . So we may as well go with the smallest possible value G = 2 .
If we have A B = 2 × 3 × 5 × 7 G = 2 then G A B = 4 2 0 which is good; and since A B has four prime factors then we can split it into a coprime pair in 2 4 = 1 6 ways - each prime factor is either in A or in B .
If we introduce a fifth distinct prime factor, the smallest possible is 1 1 but that busts G A B > 1 0 0 0 . Or if we introduce a power to the prime factors of this example, it gains nothing because of the requirement that A , B be coprime - all powers of that factor would have to stay together.
So there is no way we can expand on this solution, which is N 2 , 4 2 0 .
NB. There is another value of a b with the same N G , L value if we take G = 3 , i.e. N 3 , 6 3 0 .
Nicely done!
The end of the solution is slightly informal, though can be easily made rigorous.
One can also get the same 1 6 in a few other ways, for example by taking the pair ( 2 , 6 6 0 ) = ( 2 , 2 ⋅ 2 ⋅ 3 ⋅ 5 ⋅ 1 1 ) .
Let the prime decomposition of G be p 1 a 1 p 2 a 2 ⋯ p n a n and that of L be p 1 b 1 p 2 b 2 ⋯ p n b n , where the p i s are prime and a i s and b i s are non-negative integers. It is obvious that b i ≥ a i for all i . Note that n is the number of primes in the prime decomposition of L . If a prime p k is not present in the prime decomposition of G , then we have a k = 0 (it was pretty obvious from the definition, but I thought it is better to clarify).
Let
a
=
p
1
c
1
p
2
c
2
⋯
p
k
c
k
and
b
=
p
1
d
1
p
2
d
2
⋯
p
n
d
n
. Then, note that
g
cd
(
a
,
b
)
=
p
1
m
i
n
(
c
1
,
d
1
)
p
2
m
i
n
(
c
2
,
d
2
)
⋯
p
n
m
i
n
(
c
n
,
d
n
)
l
c
m
(
a
,
b
)
=
=
p
1
m
a
x
(
c
1
,
d
1
)
p
2
m
a
x
(
c
2
,
d
2
)
⋯
p
n
m
a
x
(
c
n
,
d
n
)
Matching powers, we obtain
m
i
n
(
c
i
,
d
i
)
=
a
i
and
m
a
x
(
c
i
,
d
i
)
=
b
i
for all
i
.
Let us determine how many possibilities this gives rise to for the
c
i
s and
d
i
s. If there are
k
i
possibilities for
(
c
i
,
d
i
)
for all
i
, then by the multiplication principle we have
N
G
,
L
=
i
=
0
∏
n
k
i
Note that if we have
a
m
=
b
m
for some
m
, then there is only one possibility for
(
c
m
,
d
m
)
:
c
m
=
d
m
=
a
m
=
b
m
This implies that
N
G
,
L
does not change if we remove all the primes from
G
and
L
that appear the same number of times in both
G
and
L
. So let us assume
a
i
=
b
i
for all
i
, thus
a
i
>
b
i
.
Note that
m
i
n
(
c
i
,
d
i
)
=
a
i
and
m
a
x
(
c
i
,
d
i
)
=
b
i
,
b
i
>
a
i
gives two possibilities for
(
c
1
,
d
i
)
, as
c
i
can take two values:
a
i
or
b
i
, and
d
i
must take the other value. Then by the multiplication principle, we have
N
G
L
=
2
n
.
We wish to find the largest possible value of n for which we can find two integers lying in the range [ 2 , 1 0 0 0 ] , of which the larger one has exactly n prime divisors, and there exists no prime which appears a same number of times in both the integers. Note that if we set G = 2 , L = 2 2 ⋅ 3 ⋅ 5 ⋅ 7 , then n = 4 and N G , L = 2 4 = 1 6 . Note that this is the largest possible value of n we can obtain, as we are trying to maximize the number of prime divisors of L while minimizing the number of prime divisors of G keeping the condition that no prime appears the same number of times in G and L . So we can conclude that our answer is 1 6 .
Note: if the 1 0 0 0 was replaced by some larger positive integer k , then our optimal choice of ( G , L ) would be ( G , L ) = ( 2 , 2 2 ⋅ 3 ⋅ 5 ⋯ p x ) , where p i is the i t h prime and x is the largest integer satisfying 2 2 ⋅ 3 ⋅ 5 ⋯ p x ≤ k .
Nice job!
For largest possible value of N G , L , observe that we would need the GCD G to be small, so that we can have more multiple-pairs of it.
Suppose we have positive integers ( a , b ) st a = G h , b = G k with ( h , k ) = 1 . Note that L = G h k .
For L ≤ 1 0 0 0 & get more pairs, the greatest as such would occur for G = 2 (& possibly for other values,but for 2 is sure,since it is the smallest prime). Then h k ≤ 5 0 0 .
To have more pairs of ( h , k ) , ( h , k ) assume values divisible by single powers of primes only(observe that by allowing any of them to be a multiple of any higher power of a prime, number of pairs will fall as h k increases but it is bounded above),
we know 2 × 3 × 5 × 7 = 2 1 0 < 5 0 0 , so we put h k = 2 ⋅ 3 ⋅ 5 ⋅ 7 , thereby furnishing 2 4 = 1 6 ordered values for ( h , k ) & in turn for ( a , b ) .
The solution is basically correct, but not very rigorous. Statements like "assume values divisible by single powers of primes only(observe that by allowing any of them to be a multiple of any higher power of a prime, number of pairs will fall as h k increases but it is bounded above" should be avoided in a rigorous proof. Instead, one should say something like: "For any pair ( G , L ) with G ∣ L , consider a pair ( G , L ′ ) , where L ′ / G is the product of all primes dividing L / G . Note that N ( G , L ′ ) = N ( G , L ) and L ′ ≤ L . Therefore, we only need to consider pairs ( G , L ) for which L / G is a product of distinct primes."
Thanks for your suggestion Challenge Master,I will try to fulfill the 'rigorous' part. I always like to 'feel' the problem & think.
Problem Loading...
Note Loading...
Set Loading...
Since G is a factor of a and a is a factor of L , so G is a factor of L , if G does not divide L , then N G , L would be 0 .
Lemma: The value of N G , L is the same as the number of pairs of positive integers ( a , b ) such that a and b are relatively prime and their Least Common Multiple is G L .
Proof: We would establish a bijection between sets A, that consists of the pairs of positive integers ( a , b ) such that G = g c d ( a , b ) and L = l c m ( a , b ) and set B, that consists of the pairs of positive integers such that a and b are relatively prime and their Least Common Multiple is G L .
We consider a certain ordered pair in A . We divide both a and b by G to get an ordered pair in B . Likewise, if we consider an ordered pair in B , we multiply both a and b by G to get an ordered pair in A . Hence we have established a bijection.
Now we find the number of pairs of positive integers ( a , b ) such that a and b are relatively prime and their Least Common Multiple is G L .
Since a and b are relatively prime, they do not share any prime factor, so for each prime factor in G L , it would exist either in a in b , but not in both, and obviously, not in none of them. Hence, there are 2 ways to choose (the exponent would be the same as the one in G L . Thus, the number of such ordered pairs is the 2 raised to the number of prime factors in G L .
Now G L is at most 5 0 0 . It could have at most 4 different prime factors because the product of the first 5 prime numbers is ( 2 ) ( 3 ) ( 5 ) ( 7 ) ( 1 1 ) = 2 3 1 0 , which is greater than 5 0 0 . Hence, the maximum possible value of N G , L is 2 4 = 1 6
Now if we let L = 4 2 0 and G = 2 , we would have G L = 2 1 0 , which has 4 different prime factors: 2 , 3 , 5 , 7 . In this case, N G , L = 2 4 = 1 6 .
Thus, the largest possible value of N G , L is 1 6 .