Adjsim

Feature development:
* Added powerful new reinforcement learning decision module: Perturbative Q-Learning (27).
* Added array-based decision-mutable values as well as variable constraint types (23):
* Boolean Array, no constraints.
* Integer Array, range constraint.
* Float Array, range and sum constraints.
* New Tutorial! (16)
* New Economic trader simulation. see examples/comparative_advantage (17).

Bugfixes: 12, 19, 21

This release marks a major architectural overhaul for AdjSim. The AdjSim framework is now class based, modular, and much cleaner than the previous release. The same simulations can be described with substantially less code. This architectural changeset entails largely equivalent functionality to the previous release.

Changes:
* Major API Architectural Overhaul. Virtually all of AdjSim was rewritten.
* Addition of simulation index semantics.
* Addition of Grid Spatial index.
* Addition of simulation callback semantics.
* Reworked demo/image/gif folder structure.
* A significantly more robust unit/system test module.

This is the proof of concept release for AdjSim.
Contained within this release are the following:
- The AdjSim Engine.
- A graphical representation of the simulation using PyQt4
- Post-simulation analysis tools implemented using matplotlib.
- The AdjSim Intelligence module, with a basic implementation of Q Learning available.
- AdjSim demos exemplifying the versatility of the engine, listed below.
- Simple bacterial behavior simulation model.
- Predator-prey relationship model.
- Jupiter moon system model exemplifying orbital mechanics and physics modelling capacities of the engine.
- A Conway's Game of Life implementation (with 2 scenarios).