High Tangle Order Gives Large Tangle-Independent Set¶
Claim/Theorem¶
Let \(M\) be a 3-connected matroid, let \(\mathcal T\) be a tangle of \(M\), let \(N\) be a minor of \(M\) with no loops or coloops, and let
in the tangle matroid \(M(\mathcal T)\). If the order of \(\mathcal T\) is at least
then there exists a set \(X\subseteq E(M)\) with
such that one of
is 3-connected with \(N\) as a minor.
Specializing to the empty minor gives:
if the tangle order is at least 20k-13, then there exists a \(\mathcal T\)-independent set \(X\) of size k such that one of \(M\backslash X\) or \(M/X\) remains 3-connected.
By Lemma 2.29 of the same paper, every \(\mathcal T\)-independent set is both independent and coindependent in \(M\). Therefore high tangle order forces a large set of elements that is simultaneously:
- independent,
- coindependent,
- and globally removable by all-deletion or all-contraction while preserving
3-connectivity.
Using [[tangle-order-equals-branchwidth.md]], the same conclusion applies to any 3-connected matroid of sufficiently large branchwidth.
Dependencies¶
- [[tangle-order-equals-branchwidth.md]]
Conflicts/Gaps¶
- This theorem produces a large flexible sparse set, not a dense or balanced highly connected set.
- It does not imply that the set \(X\) lies in a single bag of a lean tree-decomposition, nor that it yields linear cut rank across balanced separators.
- So the current frontier becomes sharper: high tangle order already gives many removable independent/coindependent elements, but the graph still lacks a theorem concentrating such a set into the kind of dense or bag-local object needed by [[dense-k-connected-set-forces-balanced-cut-rank.md]] or [[lean-matroid-bag-gives-rank-connected-set.md]].
Sources¶
10.1016/j.jctb.2014.12.003