Re:Inventing an Operation to Solve x^x = y (2000)
by fyp on Thursday November 28, 2019 @08:40PM
(#21661343)
Is this problem correctly stated? Substituting in x=f(y) you get
so
This suggests a linear solution but then you get f(x)=sqrt(3)x which doesn't work since it's N -> N.
[ Parent
| on HN ]
Re:Inventing an Operation to Solve x^x = y (2000)
by Chinjut on Thursday November 28, 2019 @10:06PM
(#21661688)
Suggests is not demands.
If we discard the requirement to be an increasing function, one recursively defined solution takes f(3x + 1) = 3x + 2, f(3x + 2) = 3 * (3x + 1), and f(3x) = 3f(x), with f(0) = 0.
Or analogously for any way of splitting the positive integers not divisible by 3 into ordered pairs (a, b). Then define f(3^n a) = 3^n b and f(3^n b) = 3^(n + 1) a. This is the general form of the solutions.