[SciPy-user] Place data on grid to minimise cost function

[John Reid]

I have a problem for which I do not know any algorithm to solve it. I'm 
wondering if there is something in scipy that will help me. Apologies if 
there isn't and this is off-topic.

I have a set of N data points, which I want to place on a k x m grid. I 
have a cost function D(x1,x2) that measures how different 2 data points 
are. I want to minimise the sum of D over all neighbouring pairs of 
points on the grid (i.e. I want similar points to be neighbours on the 
grid).

A) Is there some established technique to do this?
B) Is it in scipy?

BTW, N = k.m

Thanks,
John.