Double adding digits

Noticing that:

  • 3 6 + 3 6 = 36 3\cdot 6 + 3\cdot 6 = 36
  • 13 52 + 13 52 = 1352 13\cdot 52 + 13 \cdot 52 = 1352

Does there exist another pair of positive integers with the same number of digits satisfying: A B + A B = A B A\cdot B + A\cdot B = \overline{AB} ?

Note: The digits are in the decimal representation. A B \overline{AB} Denotes adding the two numbers as a string, for example if A = 16 , B = 1 A=16, B=1 then A B = 161 \overline{AB} = 161 .

Bonus: What if A , B A, B don't have the same number of digits? can you find such pair?

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.

2 solutions

Mustafa Alelg
Jun 2, 2018

No. ( A , B ) = ( 3 , 6 ) , ( 13 , 52 ) (A, B) = {(3, 6), (13, 52)} are the only solutions.

Proof: Assume that the number of digits in B B is n n . Then A B = 1 0 n A + B = 2 A B B ( 2 A 1 ) = 1 0 n A \overline{AB} = 10^n \cdot A + B = 2\cdot AB \Leftrightarrow B(2A-1) = 10^n \cdot A

And since g c d ( A , A 1 ) = g c d ( A , A 1 + A ) = 1 gcd(A, A-1) = gcd(A, A-1 +A) = 1 therefore 2 A 1 1 0 n 2A-1 \mid 10^n but 2 A 1 2A-1 is odd so it must be a divisor of 5 n 5^n which implies that 2 A 1 5 n A 5 n 1 2 2A-1 \leq 5^n \Leftrightarrow A \leq \frac{5^n - 1}{2}

On the other hand we know that the number of digits in both A , B A, B is equal to n n therefore A A cannot be less than 1 0 n 1 10^{n-1} and from the previous inequality we get: 1 0 n 1 A 5 n 1 2 10^{n-1} \leq A \leq \frac{5^n - 1}{2}

5 n 2 n 5 n 1 1 5 n 1 ( 5 2 n ) 1 \Leftrightarrow 5^n - 2^n \cdot 5^{n-1} \geq 1 \Leftrightarrow 5^{n-1} \cdot (5-2^n) \geq 1 and from here we find that n n cannot be big because the L H S LHS must be positive, the only possible values for n n here are only 1 and 2. From which we get the solutions mentioned above for A , B A, B .

63 504 + 63 504 = 63504 63 \cdot 504 + 63 \cdot 504 = 63504

Michael Mendrin - 3 years ago

Log in to reply

The question states that A, B has the same number of digits.

donglin loo - 3 years ago

Log in to reply

so I read poorly

Michael Mendrin - 3 years ago

Log in to reply

@Michael Mendrin Its ok i also overlooked that when answering

donglin loo - 3 years ago

That's not correct as Donglin has mentioned. But I'm amazed that you've got such a big number satisfy this, do you have a way of getting another numbers? and how did you get this number? because by this we may generalize the question.

Mustafa Alelg - 3 years ago

Log in to reply

After thinking about it, you're right, there might exist other such numbers that are even larger, where the number of digits in A , B A, B aren't the same. I used the fact that the last digit of A A has to be either 1 1 , 3 3 or 8 8 , while the last digit of B B has to be even. Then we look at the expansion

100000 A 1 + 10000 A 2 + 1000 A 3 + 100 B 1 + 10 B 2 + B 3 100000A1+10000A2+1000A3+100B1+10B2+B3
2 ( 100 A 1 + 10 A 2 + A 3 ) ( 100 B 1 + 10 B 2 + B 3 ) = 0 -2(100A1+10A2+A3)(100B1+10B2+B3)=0

and then start eliminating terms by looking at divisibility into the coefficients. There's not a whole lot of things to try. I could use this idea to speed up computer searches for larger such numbers? A straight computer search isn't practical at all for large numbers that could exist.

Michael Mendrin - 3 years ago

Log in to reply

@Michael Mendrin ...as a matter of fact, I've just now come up with

313 5008 + 313 5008 = 31345008 313 \cdot 5008+313 \cdot 5008=31345008

which I think suggests a possible pattern

Michael Mendrin - 3 years ago

Log in to reply

@Michael Mendrin Yes, there is definitely a pattern. Here's the next

1561 50008 + 1561 50008 = 156150008 1561 \cdot 50008 + 1561 \cdot 50008 = 156150008

and then after that it suddenly seems to vanish! Hmm.. could this be the largest possible such number?

Michael Mendrin - 3 years ago

Log in to reply

@Michael Mendrin aw no such luck, there's still larger such numbers

7813 500032 + 7813 500032 = 7813500032 7813 \cdot 500032 + 7813 \cdot 500032 = 7813500032
39063 5000064 + 39063 5000064 = 390635000064 39063 \cdot 5000064 + 39063 \cdot 5000064 = 390635000064

I can generate fantastically large numbers this way, but I've not proven whether there is an infinity of them. The technigue is just a bit tricky. Here's another one

76293945313 5000000000032768 + 76293945313 5000000000032768 76293945313 \cdot 5000000000032768 + 76293945313 \cdot 5000000000032768 = 762939453135000000000032768 =762939453135000000000032768

Michael Mendrin - 3 years ago

Couple more pairs: 313 * 5008 + 313 * 5008 = 3135008; 1563 * 50016 + 1563 * 50016= 156350016. ( * means "multiply")

Larisa Altshuler - 2 years, 12 months ago

Smallest example is : 3 60+3 60=360

A Former Brilliant Member - 2 years, 3 months ago
Leonel Castillo
Jun 6, 2018

Let n n be the number of digits of the numbers. Then we want to solve the diophantine equation 2 a b = 1 0 n a + b 2ab = 10^na + b

Lemma: b 0 m o d 2 n b \equiv 0 \mod 2^n

From the equation it is clear that b 0 m o d 2 b \equiv 0 \mod 2 . Now suppose b 0 m o d 2 k b \equiv 0 \mod 2^k , thus 2 a b 0 m o d 2 k + 1 2ab \equiv 0 \mod 2^{k+1} and if k < n k < n , b 0 m o d 2 k + 1 b \equiv 0 \mod 2^{k+1} . By induction, b 0 m o d 2 n b \equiv 0 \mod 2^n .

Thus, we may write b = 2 n b 0 b = 2^n b_0 and write the equation as 2 n + 1 a b 0 = 1 0 n a + 2 n b 0 2 a b 0 = 5 n a + b 0 b 0 0 m o d a 2^{n+1} a b_0 = 10^n a + 2^n b_0 \implies 2ab_0 = 5^n a + b_0 \implies b_0 \equiv 0 \mod a . so we may write b 0 = a b 1 b_0 = ab_1 . From this we can write b = 2 n a b 1 b = 2^n a b_1 . With this the equation becomes 2 a 2 b 1 = 5 n a + a b 1 2 a b 1 = 5 n + b 1 a = 5 n + b 1 2 b 1 2a^2 b_1 = 5^n a + a b_1 \implies 2ab_1 = 5^n + b_1 \implies a = \frac{5^n + b_1}{2b_1} and b = 2 n 1 ( 5 n + b 1 ) b = 2^{n-1} (5^n + b_1) .

To solve the problem, we will find a property of valid b 1 b_1 . From the equation 2 a b 0 = 5 n a + b 0 2ab_0 = 5^na + b_0 we deduce 5 n a 0 m o d b 0 5 n a 0 m o d a b 1 5 n 0 m o d b 1 5^n a \equiv 0 \mod b_0 \implies 5^n a \equiv 0 \mod ab_1 \implies 5^n \equiv 0 \mod b_1 . This means that either b 1 = 1 b_1 = 1 or 5 b 1 5 | b_1 . Considering the first case, a = 5 n + 1 2 a = \frac{5^n + 1}{2} . For n = 1 , 2 n=1,2 we find the solutions given in the problem. However, if we try n = 3 n = 3 we find that a a has 2 digits, not 3. And indeed, it turns out that 5 n + 1 2 < 1 0 n 1 \frac{5^n + 1}{2} < 10^{n-1} for n 3 n \geq 3 . Thus, for b 1 = 1 b_1 = 1 there are no other solutions than the ones given in the problem.

Now, if we assume that 5 b 1 5 | b_1 we find another contradiction. Suppose that a = 5 n + b 1 2 b 1 1 0 n 1 a = \frac{5^n + b_1}{2b_1} \geq 10^{n-1} . Then b = 2 n 1 ( 5 n + b 1 ) 2 n 1 × 1 0 n 1 2 b 1 2 n 1 0 n 1 × 5 1 0 n b = 2^{n-1} (5^n + b_1) \geq 2^{n-1} \times 10^{n-1} 2b_1 \geq 2^n 10^{n-1} \times 5 \geq 10^n . So, if a a has n n digits, b b has at least n + 1 n+1 digits. This contradicts that they must have the same amount of digits.

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...