Notions of Möbius inversion

From MaRDI portal
Publication:1932148



Abstract: M"obius inversion, originally a tool in number theory, was generalized to posets for use in group theory and combinatorics. It was later generalized to categories in two different ways, both of which are useful. We provide a unifying abstract framework. This allows us to compare and contrast the two theories of M"obius inversion for categories, and advance each of them. Among several side benefits is an improved understanding of the following fact: the Euler characteristic of the classifying space of a (suitably finite) category depends only on its underlying graph.


The original Möbius inversion as well as its generalizations by Rota and others to posets and categories all require some form of finiteness. In the article, the author introduces a unifying framework for the study of both fine and coarse Möbius inversion for suitable categories and, for the coarse version, also for enriched categories. The definitions require quite straightforward adaptations of finiteness conditions on the category, however to allow for categories with infinitely many objects, the following result about matrices, interesting on its own right, is proved in an appendix (Theorem A.4): Let \(Z\) be an invertible, transitive, \(n\times n\) matrix over \(k\). Let \(i,j\in \{1,\ldots, n\}\). Then \(Z_{ij}=0\implies (Z^{-1})_{ij}=0\). The article is very well written and is self-contained. It provides for a clear comparison between existing notions of Möbius inversion as well as significantly broader the class of structures allowing Möbius inversion, presents several interesting examples, and presents a detailed functoriality analysis.











This page was built for publication: Notions of Möbius inversion

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1932148)