The spatial package contains basic functionality for implementing spatially
embedded (or so-called "fine-grained parallel") population models.  In these
models each subpopulation has its individuals spread out in some notional
space (such as a grid), and breeding is accomplished by selecting an
individual, picking a sufficiently fit "neighbor" of that individual, breeding
the two, and replacing the original individual with the resulting child.  The 
definition of the space and the notion of neighborhood is up to the particular
space being used.  The 'Space' interface defines the necessary functions
for creating a subpopulation embedded with some space of your choosing.
Use the SpatialTournamentSelection selection method to select individuals
in your breeding pipeline, and a spatial Breeder.  Classes of interest:



ec.spatial.Space

An interface which includes methods necessary to embed notions of
some kind of space into a subpopulation.  


ec.spatial.Spatial1DSubpopulation

A basic Subpopulation which implements Space to define a 1-dimensional
toroidal space (that is, a ring).


ec.spatial.SpatialTournamentSelection

A variation of Tournament Selection which only picks neighbors of the current
individual as entrants into the tournament.


ec.spatial.SpatialBreeder

A version of Breeder which, working with SpatialTournamentSelection,
iteratively walks through the subpopulation and breeds new individuals to
replace each slot in the subpopulation while adhering to the Space constraints.


ec.spatial.SpatialMultiPopCoevolutionaryEvaluator

A multi-population coevolutionary evaluator (similar to
ec.coevolve.MultiPopCoevolutionaryEvaluator) which additionally can select 
one of the partners of a given individual(s) to be the individual in the exact
same slot in the other subpopulation(s).