![](/files/happy5.png)
Problem Does every
-connected cubic graph on
vertices admit a partition into
paths of length
?
![$ 3 $](/files/tex/4aaf85facb6534fd470edd32dbdb4e28f6218190.png)
![$ 3k $](/files/tex/7b38385e0448476b401faa5844560387ef8db56b.png)
![$ k $](/files/tex/c450c3185f7285cfa0b88d3a903c54f7df601201.png)
![$ 2 $](/files/tex/5271e36bb1c040e0f14061d89cd97d0c86d4e06f.png)
More generally, the following question is posed.
Problem Does every
-connected cubic graph on at least
vertices contain
pairwise vertex-disjoint paths of length
?
![$ 3 $](/files/tex/4aaf85facb6534fd470edd32dbdb4e28f6218190.png)
![$ 3k $](/files/tex/7b38385e0448476b401faa5844560387ef8db56b.png)
![$ k $](/files/tex/c450c3185f7285cfa0b88d3a903c54f7df601201.png)
![$ 2 $](/files/tex/5271e36bb1c040e0f14061d89cd97d0c86d4e06f.png)
In [K1], Kelmans gave a construction that provided infinitely many 2-connected graphs for which the above statement is false.
Bibliography
[K1] Alexander K. Kelmans, Packing 3-vertex paths in 2-connected graphs
*[K2] Alexander K. Kelmans, On --Packing in 3--connected Graphs, RUTCOR Research Report 23--2005, Rutgers University. See also Packing 3-vertex Paths In Cubic 3-connected Graphs
* indicates original appearance(s) of problem.