d3:
Select six weights from a set of weights (1, 2, …, 26) grams in such a way that any two subsets of the six add up to distinct total weights.
Bonus-d4: Prove that no similar subset of 7 exists.
Source: Moscow Math Olympiad
I believe the complete set of solutions are:
1 6 12 22 24 26
1 12 18 22 24 26
3 6 12 22 24 26
3 6 12 23 24 25
3 6 12 24 25 26
3 6 13 24 25 26
3 12 18 22 24 26
3 12 18 23 24 25
3 12 19 24 25 26
3 13 19 24 25 26
5 6 12 22 24 26
5 12 18 22 24 26
6 7 12 22 24 26
6 9 12 22 24 26
6 9 12 23 24 25
6 11 12 22 24 26
6 12 13 22 24 26
6 12 15 22 24 26
6 12 15 23 24 25
6 12 17 22 24 26
6 12 19 22 24 26
6 12 21 22 24 26
6 12 21 23 24 25
6 12 22 23 24 26
6 12 22 24 25 26
6 13 22 24 25 26
7 12 18 22 24 26
8 13 16 19 25 26
9 12 18 22 24 26
9 12 18 23 24 25
9 13 18 20 23 26
9 15 18 21 25 26
11 12 18 22 24 26
11 13 18 22 23 26
11 17 20 22 23 24
11 17 21 22 23 25
11 18 19 20 22 25
11 18 21 23 24 25
11 18 22 23 24 26
11 19 22 24 25 26
12 13 18 22 24 26
12 15 18 22 24 26
12 15 18 23 24 25
12 17 18 22 24 26
12 18 19 22 24 26
12 18 21 22 24 26
12 18 21 23 24 25
12 18 22 23 24 26
12 18 22 24 25 26
12 19 20 21 23 26
12 19 22 24 25 26
13 16 19 24 25 26
13 19 22 24 25 26
I didn't want to resort to a program, (just as the Olympiad youth had no access to such a resource) so I did the logical thing - experimented. As I added another weight I scratched out all the sums covered by the 2,3,4 weights I had selected. The list of sums rapidly filled-in for most sets of selected weights. (There are 63 possible subsets: 6 singletons, 15 pairs, 20, 15, 6, & 1 full sum... from binomial coefficients: 6 chose k<6. After trying to strategize, I got feed-up, threw up my hands and just programmed it. Clearly large weights help for unique sums.
My guess for showing that 7 is impossible is to show how the list of sums always completely fills. (It may be wise to start adding the heaviest weights first). While building the list of weights, all sums come back to exclude many of the remaining possibilities. In dealing with sums of sums, I catch a whiff of the Fibonacci sequence as possibly being part of an erudite proof....
Edited on January 23, 2019, 12:55 pm