Will need to implement a set of integration methods

[stiber]

Folks who want to move their simulations into BrainGrid will want to integrate the differential equations that describe synapses and/or neurons. To facilitate this, we will want to implement a small set of integrators that we have debugged, validated, and characterized. This can be done as part of a future step in which we are simulating other models as demonstrations. I suggest the minimal set of methods would be one explicit method (such as Runge-Kutta) and one implicit method.