A remark on partial sums involving the Mobius function
Abstract
Let < > ⊂ be a multiplicative subsemigroup of the natural numbers = \1,2,3,...\ generated by an arbitrary set of primes (finite or infinite). We given an elementary proof that the partial sums Σn ∈ < >: n ≤ x μ(n)n are bounded in magnitude by 1. With the aid of the prime number theorem, we also show that these sums converge to Πp ∈ (1 - 1p) (the case when is all the primes is a well-known observation of Landau). Interestingly, this convergence holds even in the presence of non-trivial zeroes and poles of the associated zeta function ζ(s) := Πp ∈ (1-1ps)-1 on the line \(s)=1\. As equivalent forms of the first inequality, we have |Σn ≤ x: (n,P)=1 μ(n)n| ≤ 1, |Σn|N: n ≤ x μ(n)n| ≤ 1, and |Σn ≤ x μ(mn)n| ≤ 1 for all m,x,N,P ≥ 1.
Turn this paper into a lesson
ArcXiv compiles a structured reading guide from this paper's metadata: plain-English importance, contributions, prerequisite concepts, which sections to read first, flashcards, and a quiz. Grounded in the abstract, never invented.