Bang–Jensen–Gutin Prescribed-Endpoint Path Problems

Given a semicomplete digraph DD and distinct vertices x,yV(D)x,y\in V(D), determine a longest directed (x,y)(x,y)-path. Given a locally semicomplete digraph DD and distinct vertices x,yV(D)x,y\in V(D), determine whether DD contains a Hamiltonian directed (x,y)(x,y)-path, that is, a directed path from xx to yy containing every vertex of DD exactly once. The general problem remains unresolved for strong, nonsemicomplete, non-44-strong locally semicomplete digraphs.

Progress summary

Partially solved

A 2026 paper settles several important special cases, but the full prescribed-endpoint path problem remains open for a broad class of directed networks.

The problem asks when directed graphs contain Hamiltonian or longest paths joining specified endpoints. The recent work by Bai, Hanzhi, Yan, Jin, Zhou, and Jia advances both questions through structural reductions.

August 2026 structural progress

The paper settles several structural cases in semicomplete and locally semicomplete digraphs and gives an equivalent cycle formulation. It explicitly leaves the strong, nonsemicomplete, non-44-strong case open.

Current status (as of August 2026): Several semicomplete and locally semicomplete cases are settled, while strong nonsemicomplete locally semicomplete digraphs remain unresolved.

Sources
Sources & referencesView supporting material

Primary source

arXiv

Additional references

Solutions 0

No solutions have been posted yet.