Four-page index and linear upper bounds for ribbonlength
Hyungkee Yoo
TL;DR
This work links geometric ribbonlength to a new diagrammatic invariant, the four-page index $\alpha_4(K)$, by leveraging a Kauffman state determined from spanning trees to produce a binding circle and a four-page presentation. It proves the central bound $\mathrm{Rib}(K) \le \alpha_4(K)$ and establishes $\alpha_4(K) \le 2c(K)$, with strict inequality for non-alternating links, yielding the linear bound $\mathrm{Rib}(K) \le 2c(K)$ and providing a practical diagrammatic method to estimate ribbonlength. The paper also develops circular four-page presentations via binding circles and demonstrates sharpness of the coefficient, alongside concrete examples. Overall, the results give a sharper, diagrammatic linear bound for ribbonlength and connect open-book presentations to quantitative knot complexity.
Abstract
We introduce the four-page index of a knot or link as a presentation invariant arising from embeddings in a four-page open book decomposition. Using spanning trees of the checkerboard graph of a reduced non-split diagram, we construct a Kauffman state consisting of a single state circle. The associated Eulerian tour of the underlying 4-valent plane graph determines a binding circle intersecting each edge exactly once, producing a four-page presentation with at most $2c(K)$ arcs. Hence $$ α_4(K) \le 2c(K), $$ with strict inequality in the non-alternating case. We further prove that ribbonlength is bounded above by the four-page index, and therefore obtain the linear bound $$ \mathrm{Rib}(K) \le 2c(K). $$ This improves the previously known general linear upper bound for ribbonlength and provides a diagrammatic method for estimating ribbonlength.
