Maths Olympiad Prep

Library / /73 of 87

Number theory Difficulty 6.7 National Olympiad Prove it Austria

The equation n!+An=nkn! + A \cdot n = n^k has (n,k)=(0,0)(n, k) = (0, 0) as a solution for every non-negative integer AA. Determine all non-negative integer solutions of this equation for A=7A = 7 and A=2012A = 2012.

Solution

We first note that, independently of the value of AA, n=0n = 0 implies k=0k = 0 and vice versa. Furthermore, also independently of the value of A>0A > 0, there can be no solution for either n=1n = 1 or k=1k = 1, since n!+An>nkn! + A \cdot n > n^k certainly holds in either of these cases. In the following, we therefore limit our discussion to the case n,k2n, k \ge 2. In this case we can divide the given equation by nn, which yields the equivalent equation
(n1)!+A=nk1. (n - 1)! + A = n^{k-1}.
We now turn our attention to the case A=7A = 7.
If n=2n = 2, the equation yields 1+7=2k11 + 7 = 2^{k-1}, which is true for k=4k = 4, yielding the solution (n,k)=(2,4)(n, k) = (2, 4). If n>2n > 2, (n1)!(n - 1)! is even, and nk1n^{k-1} and therefore nn must be odd. If n=3n = 3, we have (31)!+7=331(3 - 1)! + 7 = 3^{3-1}, and we see that (n,k)=(3,3)(n, k) = (3, 3) is another solution. We now wish to show that there are no others.
For n{5,7,11,13}n \in \{5, 7, 11, 13\} Wilson's theorem yields (n1)!+71+76(modn)(n - 1)! + 7 \equiv -1 + 7 \equiv 6 \pmod{n}, which means that this expression cannot be a power of nn. For n=9n = 9 we obtain 09k18!+71(mod3)0 \equiv 9^{k-1} \equiv 8! + 7 \equiv 1 \pmod{3}, which again yields a contradiction. Finally, for n15n \ge 15 we note that 0(n1)!+7nk1(mod7)0 \equiv (n - 1)! + 7 \equiv n^{k-1} \pmod{7}, which implies 7n7|n. Since we obviously also have 72(n1)!7^2|(n - 1)! in this case, it follows that 0nk1(n1)!+77(mod7)0 \equiv n^{k-1} \equiv (n - 1)! + 7 \equiv 7 \pmod{7}, which once more yields a contradiction.
The only case left to check is k=2k = 2, but this yields the equation (n1)!+7=n(n - 1)! + 7 = n, which cannot hold because of (n1)!+7(n1)+7>n(n - 1)! + 7 \ge (n - 1) + 7 > n, and we see that the only two non-trivial solutions in this case are (2,4)(2, 4) and (3,3)(3, 3).

Now, let us consider the case A=2012A = 2012. We will show that there are no non-trivial solutions in this case.
For n=2n = 2 the equation reduces to 2013=2k12013 = 2^{k-1}, which obviously has no integer solution. For n>2n > 2, we have 2(n1)!2|(n - 1)!, and therefore nk1=(n1)!+20120(mod2)n^{k-1} = (n - 1)! + 2012 \equiv 0 \pmod{2}, which implies n0(mod2)n \equiv 0 \pmod{2}. n=4n = 4 yields 2018=4k12018 = 4^{k-1} for the equation, which again has no integer solution. We therefore have only the case n6n \ge 6 left to consider.
If k>3k > 3, the fact that nn is even yields 0nk1(n1)!+20124(mod8)0 \equiv n^{k-1} \equiv (n - 1)! + 2012 \equiv 4 \pmod{8}, which is a contradiction. For k=2k = 2, the equation (n1)!+2012=n(n - 1)! + 2012 = n cannot have a solution because (n1)!+2012>n1+2012>n(n - 1)! + 2012 > n - 1 + 2012 > n certainly holds. The only case left is therefore k=3k = 3, but this case does not yield a solution either, since (n1)!+2012=n2(n - 1)! + 2012 = n^2 contradicts
(n1)!+2012>(n1)(n2)(n3)+2012=n36n218+2012>n2 (n - 1)! + 2012 > (n - 1)(n - 2)(n - 3) + 2012 = n^3 - 6n^2 - 18 + 2012 > n^2
(which is true because it is equivalent to n2(n7)+12n+1994>0n^2(n - 7) + 12n + 1994 > 0).
We see that there is no non-trivial solution for A=2012A = 2012, as claimed. \square

Want a route through all this instead of an archive? The track puts 2,000 problems in a working order, from AMC 10 level to the IMO shortlist.

Source: MathNet, licensed CC-BY-4.0. Statement reproduced verbatim; metadata (topic, difficulty) added by this project.