Jamison’s questions on the number and average length of subpaths

For every connected graph GG with nn vertices, let P(G)\mathcal{P}(G) be the set of its simple subpaths, including the one-vertex paths, and define ℓ‾(G)=1∣P(G)∣∑P∈P(G)∣V(P)∣\overline{\ell}(G)=\frac{1}{|\mathcal{P}(G)|}\sum_{P\in\mathcal{P}(G)}|V(P)|. Then ℓ‾(G)≤ℓ‾(Kn)\overline{\ell}(G)\leq\overline{\ell}(K_n); equivalently, the complete graph maximizes the average length of a subpath among connected graphs of order nn.

References

Primary source

arXiv

Progress summary

Refreshed
Claimed solved

A new preprint reportedly answers the questions, but its claim that the complete graph is best has not been independently verified.

Jamison’s questions concern extremal counts and average lengths of subpaths in graphs. The retrieved sources connect the minimum-average question to Jamison’s tree result and later conjectures, but do not state every original formulation.

Known results

  • Jamison, 1984: the path gives the minimum average connected-subgraph order for trees.
  • Kroeker, Mol, and Oellermann, 2018: conjectured the same minimum for all connected graphs.
  • Two 2021 preprints proved, for connected graphs of order nn, A(G)≥(n+2)/3A(G)\geq (n+2)/3, with equality only for PnP_n.
  • A 2025 preprint records the KnK_n maximum as conjectural and discusses flawed approaches.

August 24, 2026 claimed resolution

A dated report says the preprint The number and average length of subpaths in graphs proves the analogous average-length questions and that KnK_n maximizes average path length. This is a claimed complete resolution, but the retrieved evidence does not independently verify the proof or reconcile it with the 2025 preprint’s statement that the KnK_n maximum remained conjectural.

Current status (as of August 2026): The minimum-average result is established, while the claimed proof that KnK_n maximizes average path length and answers all of Jamison’s questions remains unverified.

Sources

Solutions 0

No solutions have been posted yet.