Implements an oscillator neural network which uses a genetic algorithm as its 
learning/training rule. Flexible enough to accomodate any network structure 
and to train by weights or by varying inherent properties of the oscillators.

The primary purpose of this code is to simulate and observe this type of 
neural network (one made of oscillating nodes with various connections).
The possible applications remain to be seen, and the author's primary
goal is to investigate network architectures and initial conditions
which cannot be described analytically. Most of the current work
on these networks uses simple arrangements (chains, lattices, global networks)
and puts tight constraints on the initial conditions in order to 
obtain analytical descriptions. The author believes that there needs
to be a way to study networks which do not fall into this category
and thinks that a numerical simulation such as this implementation
is a good start.