Find the largest possible integer value of N such that g cd ( N , 2 0 3 3 ) = 1 and 2 0 3 3 N cannot be written as the sum of two positive rational numbers, each with denominators strictly less than 2033.
Details and assumptions
As an explicit example, since 2 0 3 3 1 7 0 = 2 1 4 1 + 3 8 3 , hence the answer is not 170.
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.
I first proved that the two fractions were a/107k and b/19k, then used brute-force
How do you prove that the first addend was 1/19j? How could you prove the numerator is 1 and nothing else?
Good solution! How much time did you spend in this question? looks long...
The more general statement is that if 0 < p < q are primes, then
Claim: p q N can be written as the sum of 2 positive rationals with denominators strictly less than p q if and only if there exist integers k , a , b such that
1 ≤ k ≤ p , a ≥ 1 , b ≥ 1 , N k = q a + p b
As such, the answer would be N = q − p .
So first I restated the problem: N z = 1 9 x + 1 0 7 y , 1 < = z < = 1 8 , x > 0 , y > 0 To change the question into this, first set b a + d c = 2 0 3 3 N , where ( a , b ) and ( c , d ) are each coprime. Now I will prove that b*d is a multiple of 2033. If it is not, then by adding the two fractions together with b*d as denominator and setting it equal to N / 2 0 3 3 , N must not be coprime to 2033. Now, using the coprimality of a and b, you can show that b is a factor of d. Using the coprimality of c and d, you can show that d is a factor of b. That proves b=d, which then will reduce to my restatement above, where x=a,y=c,and z=b and d. N cannot be more than 113 because then setting z=18 z N is more than 2033, which leads to there being at least 19 options for y, at least one of which must lead to a solution.Then trial and error lead me to 88. I know this isn't so clear, but it's how I did it.
Interesting solution! May I clarify how you "us[ed] the coprimality of a and b, [to] show that b is a factor of d", since that condition does not hold in the example given in the question ( 2 1 4 1 + 3 8 3 = 2 0 3 3 1 7 0 )...
Log in to reply
I meant that setting 1 9 b a + 1 0 7 d c = 2 0 3 3 N leads to my conclusion. My mistake. If b ∗ d is a multiple of 2033, then one must be a multiple of 19 and the other of 107, otherwise they cannot both be lower than 2033. I left out that step.
Problem Loading...
Note Loading...
Set Loading...
Suppose that N ∈ N with ( N , 2 0 3 3 ) = 1 . Note that 2 0 3 3 = 1 9 × 1 0 7 . Since Z / 1 9 Z is a field, we can find 1 ≤ j ≤ 1 8 such that N j ≡ 1 2 ≡ 1 0 7 modulo 1 9 .
If N ≥ 1 0 8 then N j − 1 0 7 ∈ N is congruent to 0 modulo 1 9 , and hence we can write N j = 1 0 7 + 1 9 c for some c ∈ N . Thus 2 0 3 3 N = 1 9 j 1 + 1 0 7 j c and (since 1 ≤ j ≤ 1 8 ) we deduce that 1 ≤ 1 9 j , 1 0 7 j < 2 0 3 3 .
If N ≥ 5 4 and N ≡ 1 2 modulo 1 9 , then we must have j ≥ 2 . But then N j − 1 0 7 ∈ N is congruent to 0 modulo 1 9 , and hence we can write N j = 1 0 7 + 1 9 c for some c ∈ N . Thus 2 0 3 3 N = 1 9 j 1 + 1 0 7 j c and, again, 1 ≤ 1 9 j , 1 0 7 j < 2 0 3 3 .
The numbers 5 4 ≤ N ≤ 1 0 7 which are congruent to 1 2 modulo 1 9 are 6 9 , 8 8 , 1 0 7 , and 1 0 7 is not coprime to 2 0 3 3 . Thus we know that, for any integer N ≥ 8 9 which is coprime to 2 0 3 3 , 2 0 3 3 N can be written in the desired form. It remains to confirm that 8 8 cannot be dealt with similarly.
Suppose that we can write 2 0 3 3 8 8 = b a + d c where a , b , c , d ∈ N , ( a , b ) = ( c , d ) = 1 and 1 ≤ b , d < 2 0 3 3 . Then 8 8 b d = 2 0 3 3 ( a d + b c ) , and hence both 1 9 and 1 0 7 divide 8 8 b d , and hence both divide b d . Since b , d < 2 0 3 3 , we deduce that 1 9 must divide one of b , d , with 1 0 7 dividing the other. Without loss of generality, assume that 1 9 divides b , and that 1 0 7 divides d , so that b = 1 9 u and d = 1 0 7 v where u , v ∈ N are such that ( a , 1 9 u ) = ( c , 1 0 7 v ) = 1 and 1 ≤ u < 1 0 7 and 1 ≤ v < 1 9 . Then 8 8 u v = 1 0 7 a v + 1 9 b u . From this we deduce that u ∣ 1 0 7 a v . Now 1 ≤ u < 1 0 7 , so ( u , 1 0 7 ) = 1 . Moreover ( u , a ) = 1 . Thus we deduce that u ∣ v . Similarly v ∣ 1 9 b u . Now 1 ≤ v < 1 9 , so ( v , 1 9 ) = 1 . Moreover ( v , b ) = 1 . Thus we deduce that v ∣ u . Hence it follows that u = v , so that 8 8 u = 1 0 7 a + 1 9 b where 1 ≤ u < 1 9 . Then 1 2 u ≡ 8 8 u ≡ 1 0 7 a ≡ 1 2 a ( 1 9 ) so we deduce that u ≡ a modulo 1 9 . From this it is clear that a ≥ u , and hence that 1 0 7 a + 1 9 b > 1 0 7 a > 8 8 u which is impossible. Thus we deduce that 2 0 3 3 8 8 cannot be expressed in this way, and hence the answer to the question is 8 8 .
A computer search showed me the need for one of the two fractions to be of the form 1 9 j 1 ; once I had that to aim for, the rest fell into place.