Monoidal Width

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Di Lavore, Elena, Sobociński, Paweł
Format: Preprint
Published: 2022
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916122735935488
author Di Lavore, Elena
Sobociński, Paweł
author_facet Di Lavore, Elena
Sobociński, Paweł
contents We introduce monoidal width as a measure of complexity for morphisms in monoidal categories. Inspired by well-known structural width measures for graphs, like tree width and rank width, monoidal width is based on a notion of syntactic decomposition: a monoidal decomposition of a morphism is an expression in the language of monoidal categories, where operations are monoidal products and compositions, that specifies this morphism. Monoidal width penalises the composition operation along ``big'' objects, while it encourages the use of monoidal products. We show that, by choosing the correct categorical algebra for decomposing graphs, we can capture tree width and rank width. For matrices, monoidal width is related to the rank. These examples suggest monoidal width as a good measure for structural complexity of processes modelled as morphisms in monoidal categories.
format Preprint
id arxiv_https___arxiv_org_abs_2212_13229
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Monoidal Width
Di Lavore, Elena
Sobociński, Paweł
Logic in Computer Science
Category Theory
We introduce monoidal width as a measure of complexity for morphisms in monoidal categories. Inspired by well-known structural width measures for graphs, like tree width and rank width, monoidal width is based on a notion of syntactic decomposition: a monoidal decomposition of a morphism is an expression in the language of monoidal categories, where operations are monoidal products and compositions, that specifies this morphism. Monoidal width penalises the composition operation along ``big'' objects, while it encourages the use of monoidal products. We show that, by choosing the correct categorical algebra for decomposing graphs, we can capture tree width and rank width. For matrices, monoidal width is related to the rank. These examples suggest monoidal width as a good measure for structural complexity of processes modelled as morphisms in monoidal categories.
title Monoidal Width
topic Logic in Computer Science
Category Theory
url https://arxiv.org/abs/2212.13229