Find the largest positive integer x such that x is divisible by all the positive integers at most equal to 3 x .
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.
Let's push your current ideas even further. For clarity, define λ x = L C M ( 1 , 2 , … , ⌊ 3 x ⌋ ) .
Suppose there exists an
x
≥
5
1
2
such that
λ
x
∣
x
.
Since
x
≥
5
1
2
, this implies that
λ
8
∣
x
, so
8
4
0
∣
x
and thus
x
≥
8
4
0
.
Since
x
≥
8
4
0
, this implies that
λ
9
∣
x
, so
2
5
2
0
∣
x
and thus
x
≥
2
5
2
0
.
Since
x
≥
2
5
2
0
, this implies that
λ
1
3
∣
x
, so
3
6
0
3
6
0
∣
x
and thus
x
≥
3
6
0
3
6
0
.
We would like for this to continue. For that to happen, we need additional primes, or prime powers, that that will occur in the gap of
3
2
5
2
0
to
3
3
6
0
3
6
0
, so on and so forth.
Hint: Bertrand's postulate states that there is a prime number between n and 2 n .
Problem Loading...
Note Loading...
Set Loading...
I figured a a few times wrongly on this one before getting it correct; a very interesting problem that I am still having trouble proving!
For any such x ∈ N , we must have that the least common multiple of the set of numbers { 1 , … , ⌊ 3 x ⌋ } , call it λ , divides x . Note that λ is constant on the intervals between the perfect cube numbers.
After running some pretty convincing numerical results, I found that λ > x for x ≥ 8 3 = 5 1 2 , thus contradicting λ ∣ x . It would then follow that the interval [ 3 4 3 , 5 1 2 ] , where λ = 4 2 0 , would contain our maximum value x = 4 2 0 .
In order to rigorously prove the numerical claim, I'd like to find a way to describe the growth of λ . It is a piecewise constant function of x , with discontinuities at each x = p 3 n , where n ∈ N and p is prime. Below are the graphs of y = λ ( x ) in blue and y = x in red (with logarithmic scaling for both functions on the y-axis).
Domain is [ 0 , 1 0 ] :
Domain is [ 0 , 2 0 ] :
Domain is [ 0 , 5 0 ] :
The "accumulated lcm" is apparently growing much quicker than the identity function, but it seems like proving this would have something to do with the distribution of the primes over N . If someone could give me a gentle hint in the right direction, that would be lovely!