logo

Mobius Inversion Formula Derivation 📂Number Theory

Mobius Inversion Formula Derivation

Formulas 1

ff and gg are arithmetic functions and μ\mu is the Möbius function. f(n)=dng(d)    g(n)=dnf(d)μ(nd) f(n) = \sum_{d \mid n} g(d) \iff g(n) = \sum_{d \mid n} f(d) \mu \left( {{ n } \over { d }} \right)

Explanation

The Möbius function might seem unnatural at first glance, but it actually appears in the core formulas that permeate all arithmetic functions. An arbitrary arithmetic function gg’s series can be expressed using the unit function uu and convolution \ast as gug*u, precisely because the inverse of uu is μ\mu, in other words, I=μ uI = \mu \ast\ u. Strategy: uu is the unit function and II is the identity. The series relationship of arithmetic functions with convolution can be easily derived.

Derivation

()(\Rightarrow)

f(n)=dng(d)=dng(d)u(nd) f(n) = \sum_{d \mid n} g(d) = \sum_{d \mid n} g(d) u \left( {{ n } \over { d }} \right) can be expressed as a convolution product, which is f=g uf = g \ast\ u.

Möbius series: I(n)=dnμ(d)=dnμ(d)u(nd) I(n) = \sum_{d \mid n } \mu (d) = \sum_{d \mid n } \mu (d) u \left( {{ n } \over { d }} \right)

Likewise, the Möbius series expressed as a convolution product is I=μ uI = \mu \ast\ u therefore f μ=(g u) μ=g (u μ)=g I=g \begin{align*} f \ast\ \mu =& (g \ast\ u) \ast\ \mu \\ =& g \ast\ ( u \ast\ \mu ) \\ =& g \ast\ I \\ =& g \end{align*}


()(\Leftarrow)

f μ=g    f μ u=g u    f I=g u    f=g u    f(n)=dng(d)u(nd)=dng(d)\begin{align*} & f \ast\ \mu = g \\ \implies& f \ast\ \mu \ast\ u = g \ast\ u \\ \implies& f \ast\ I = g \ast\ u \\ \implies& f = g \ast\ u \\ \implies& f(n) = \sum_{d \mid n} g(d) u \left( {{ n } \over { d }} \right) = \sum_{d \mid n} g(d) \end{align*}


  1. Apostol. (1976). Introduction to Analytic Number Theory: p32. ↩︎