1 ≤ x , y , z ≤ 1 0 ∑ min ( x , y , z ) = ?
Clarification : ( x , y , z ) is an ordered triple of integers.
Bonus : The answer is a Stirling number; coincidence?
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.
Nice approach to count this!
Yes, that's a very nice way to think about it (and a great figure)! (+1)
I wonder whether this method works here ;)
How did you approach the 3rd step?
We can show that 1 ≤ x , y , z ≤ N ∑ min ( x , y , z ) = k = 1 ∑ N k 3 = ( 2 N ( N + 1 ) ) 2
For N = 1 0 this gives 3 0 2 5 .
By induction, it suffices to show that ∑ max ( x , y , z ) = N min ( x , y , z ) = N 3 . Indeed, paying attention to the number of N ′ s in ( x , y , z ) , we find this sum to be N + 3 ∑ k = 1 N − 1 k + 3 ∑ k = 1 N − 1 k 2 = N 3 .
(See my alternative solution here .)
Very interesting problem. I divided the sum according to the value of min (x,y,z). If min (x,y,z) is "n", then it is because only one of the three numbers is n [ (3(10-n)^2) cases], two of them are n [3(10-n) cases] or the three of them are n [1 case]. If we multiply the number of cases were min (x,y,z) is n, by the value of n, and we sum it for n=1 to n=10, we get the solution :)
Yes that's exactly how I did it initially! Thanks! (+1)
Now 3025 is the Stirling Number of the Second Kind S ( 9 , 3 ) ... coïncidence?
Log in to reply
It is a coincidence, because if you replace 10 with 9 in your problem, then the answer is not S(8,3)
Log in to reply
I guess it's the French spelling ; I did not realize that it's not used in English (remember, English is my third language). The word is not pronounced like "coin", after all...
Log in to reply
@Otto Bretscher – Dang it. I thought I'm about to crack one of your weird obscure hints.
@Otto Bretscher – In Dutch one would write "coïncidentie", except we prefer "toevalligheid".
I did not have time to pursue it but it seems that the equation could be "easily" generalised. Perhaps when doing so, a relation to the Stirling numbers can be substantiated and even explained. Have you seen any more values of "n" for which that is true?
Problem Loading...
Note Loading...
Set Loading...
I ended up with the same calculation as Otto here, but used a geometric model, thinking of the numbers to be added as elements in a 1 0 × 1 0 × 1 0 grid.
Consider a set of nested cubes:
cube 1 is the entire 1 0 × 1 0 × 1 0 grid;
cube 2 is the 9 × 9 × 9 grid consisting of all elements with x , y , z ≥ 2 ;
cube 3 is the 8 × 8 × 8 grid consisting of all elements with x , y , z ≥ 3 ;
etc. until cube 10, which consists only of ( 1 0 , 1 0 , 1 0 ) .
It is not difficult to see that min ( x , y , z ) is equal to the number of cubes in which ( x , y , z ) is contained. Therefore x , y , z ∑ min ( x , y , z ) = x , y , z ∑ # of cubes containing ( x , y , z ) = n = 1 ∑ 1 0 elements contained in cube n = n = 1 ∑ 1 0 n 3 = ( n = 1 ∑ 1 0 n ) 2 = 5 5 2 = 3 0 2 5 .