Cranston–Li–Su–Wang–Xu conjecture on connectivity of nowhere-zero flows
Let be a graph in which every vertex has even degree, and let be a finite abelian group with . Define an -flow on by orienting every edge and assigning to each edge a nonzero element of , subject to flow conservation at every vertex. A cycle change consists of choosing a cycle and an element , then adding to the values of edges oriented along the cycle and subtracting from the values of edges oriented against the cycle, with the resulting flow required to remain nowhere-zero. The conjecture asserts that any two nowhere-zero -flows of can be joined by a finite sequence of such cycle changes through nowhere-zero -flows. It also asserts that, for every integer , any two integer flows on satisfying for every edge can be joined by a finite sequence of cycle changes, where each change adds a nonzero integer to the values along a cycle and all intermediate flows satisfy .
References
- The Cranston--Li--Su--Wang--Xu (2026) conjecture is true — arXiv — Dave Goel
Source: arXiv
Progress summary
The conjecture remains open: a 2026 paper proves only special cases, and no independently checked complete solution was found.
The conjecture, posed by Daniel W. Cranston, Jiaao Li, Bo Su, Zhouningxin Wang, and Ningyan Xu in 2026, concerns connectivity of reconfiguration graphs of nowhere-zero flows. It includes the proposed equivalence between integer-valued and finite-abelian-group formulations.
Known results
- Cranston, Li, Su, Wang, and Xu (2026): every Eulerian graph is -flow-connected for every positive integer .
- Cranston, Li, Su, Wang, and Xu (2026): the conjectured Eulerian-graph statement is established when or is even.
- Esperet et al. constructed graphs where is connected but is disconnected.
2026 partial progress
Cranston, Li, Su, Wang, and Xu explicitly state that the reverse implication in their central equivalence remains open; their paper presents the odd cases and the full conjecture as unresolved. No retrieved source provides independent mathematical assessment of a complete proof.
Current status (as of November 2026): special cases and a partial connectivity theorem are known, but the central equivalence and the remaining odd cases remain open; no verified complete resolution is recorded.
Sources
Solutions 0
No solutions have been posted yet.