Skip to content

High Tangle Order Gives Large Connected Set

Claim/Theorem

Let \(M\) be a 3-connected matroid with a tangle of order \(\theta\). Then for every integer \(m\) with

\[ \theta\ge 20m-13, \]

\(M\) contains a connected set \(X\) of size m satisfying

\[ \lambda_M(A)\;\ge\;\min\{|A\cap X|,\ |X-A|\} \]

for all \(A\subseteq E(M)\).

Equivalently, \(X\) is an (m+1)-connected set in the sense of [[tangle-breadth-gives-k-connected-set.md]].

Proof:

  1. By [[high-tangle-order-gives-large-tangle-independent-set.md]], under the same order hypothesis there exists a \(\mathcal T\)-independent set \(X\) of size m.
  2. By [[tangle-independent-set-gives-connected-set.md]], any such \(X\) is a connected set.

Using [[tangle-order-equals-branchwidth.md]], the same statement can be rephrased in branchwidth language:

if a 3-connected matroid has branchwidth bw(M), then it contains a connected set of size

\[ \Omega(\operatorname{bw}(M)). \]

This is a real upgrade over the earlier intrinsic route: high width no longer gives only an abstract large tangle or a removable independent/coindependent set, but an actual cardinality-based connected set in the original 3-connected matroid.

Dependencies

  • [[high-tangle-order-gives-large-tangle-independent-set.md]]
  • [[tangle-independent-set-gives-connected-set.md]]
  • [[tangle-order-equals-branchwidth.md]]

Conflicts/Gaps

  • The connected set obtained here has size only Omega(branchwidth). For good codes this is still only Omega(n/log n), not a dense linear fraction of the ground set.
  • The theorem requires 3-connectivity. For Conjecture 3 this is currently accessed only after passing to the weakly 4-connected minor route, so there remains a minor-versus-original-matroid gap.
  • Therefore the new theorem materially strengthens the intrinsic route but still falls short of the dense linear connected set needed to close the separator-alignment gap outright.

Sources

  • 10.1016/j.jctb.2014.12.003
  • 10.37236/12467