![](/files/happy5.png)
The three 4-flows conjecture
![$ G $](/files/tex/b8e7ad0330f925492bf468b5c379baec88cf1b3d.png)
![$ A_1,A_2,A_3 \subseteq E(G) $](/files/tex/00b3c9754dad6733bd2accc303abd67548500b18.png)
![$ A_1 \cup A_2 \cup A_3 = E(G) $](/files/tex/9221eff05fcc813555f4a22eb3ce690354664239.png)
![$ G \setminus A_i $](/files/tex/880e1ceffd87347d797ae0d4e59d524734f42f73.png)
![$ 1 \le i \le 3 $](/files/tex/f40f4d8f17c01da8627f69cc31a03ec8efa58853.png)
A graph has a nowhere-zero 4-flow if and only if there exist disjoint sets
with
so that
has a nowhere-zero 2-flow for
. Thus, the above conjecture is true with room to spare for such graphs. Since every 4-edge-connected graph and every 3-edge-colorable cubic graph has a nowhere-zero 4-flow, this conjecture is automatically true for these families. As with the 5-flow conjecture or the cycle double cover conjecture, establishing this conjecture comes down to proving it for cubic graphs which are not 3-edge-colorable.
This conjecture is a consequence of the Petersen coloring conjecture, and it implies the Orientable cycle four cover conjecture. The latter implication follows immediately from the fact that every graph with a nowhere-zero 4-flow has an orientable cycle double cover. Actually, it is possible that for every graph with no cut-edge, there exist disjoint sets
with
and so that
and
have nowhere-zero 3-flows and
has a nowhere-zero 2-flow. The Petersen graph has such a decomposition (
and
should be alternate edges of some 8-circuit) and so does every graph with a nowhere-zero 4-flow. If this stronger statement is true, then it would imply the oriented eight cycle four cover conjecture.
Bibliography
[J] F. Jaeger, On circular flows in graphs. Finite and infinite sets, Vol. I, II (Eger, 1981), 391--402, Colloq. Math. Soc. János Bolyai, 37, North-Holland, Amsterdam, 1984.. MathSciNet
* indicates original appearance(s) of problem.