Additional ideas...
You can epoxy one round barrel magnet flush with the end of a short metal tube.
Take the platen off the end of the screw of a C-Clamp, leaving the ball end only.
Then put enough epoxy into the tube to hold a second magnet,
and insert it so the poles your pushing together are the same (Repelling NN or SS).
Screw the clamp down until the magnets touch, an wait for the epoxy to dry FULLY (LOL).
I'd wait two full days, not just the 24-hours the package will state...
Then cut off the unused remainder of the metal tube.
You can use a ferrous (IE: steel) tube to provide a closed flux path
from the center of both magnets out to the ends of each magnet.
Picture the H-frame of a transformer, with the magnets the center pole,
but instead of having two outside poles to provide a return flux path,
you have a completely encasing tubular return flux path instead.
The above is an old old idea,
but there is a variant that allows the return path to be turned on and off.
Just use a nonferrous tube (IE: copper) instead of steel,
next wind two coils starting at the tube/magnets center,
(or a single center-tapped wire) and wind to both ends.
Starting from the center, wind both coils in the same direction onto the tube,
so the appearance is one is wound clockwise, the other counter-clockwise.
When you induce a flux into the copper from the coils,
the copper will also include the flux from the magnets,
and the effect will cease when the coil's power does.
I'm at the risk of pointing out the obvious here perhaps,
but the center tap (or the two wire ends at the middle)
get one polarity in, and the two ends get the other polarity.
You could also do this without the coils be flowing current through the tubing,
one polaritiy applied to the center of the tube right where the magnets meet,
the other polarity applied to both ends of the copper tubing piece equally.
But this would require an insane amount of current,
so as such, this is a completely inviable method.
Something similar to this is how magnetic locking doors work
as an application example for you to judge concept eligibility.
Either way, you wind up with the same pole at each ends,
for whatever reason you think this will help your concept.
You might want to read up on "Parallel-Path Magnetics" too...
http://www.google.com/search?q=Magnet+%22Parallel+Path%22&hl=en&newwindow=1&gbv=2&num=100&lr=lang_en&ft=i&cr=&safe=off