3 ms·
The Riemann hypothesis is equivalent to a number of interesting statements. One that I like is this. The Möbius function, defined by http://mathworld.wolfram.
by btilly 10y ago
The Riemann hypothesis is equivalent to a number of interesting statements. One that I like is this.
The Möbius function, defined by http://mathworld.wolfram.com/MoebiusFunction.html http://mathworld.wolfram.com/MoebiusFunction.html, shows up in various sieving operation. It is 0 for a random integer with probability (1 - 6/pi^2) and otherwise has apparently even odds of being +-1.
Its sum is called Mertens function, see http://mathworld.wolfram.com/MertensFunction.html http://mathworld.wolfram.com/MertensFunction.html. The Riemann conjecture is equivalent to saying that the growth of Mertens function is o(n^(0.5+e)) for every e > 0. For a long time it was thought to be bounded above by n^0.5, but this is known to be wrong.
If you replace the Möbius function with a function that is -1, 0 and 1 with the right probabilities, then with probability 1 it is O((n * log(log(n))^0.5). Which would prove the Riemann hypothesis. (This intuition would have kept Mertens from conjecturing that it is bounded by +-sqrt(n)...)