Add behavior mode identification

[JamesPHoughton]

In the modeling process, we often specify a behavior mode, and then want to see if our model replicates that mode. With more nuance, we might want to find the regions of the parameter space where a behavior mode is dominant. This would be nice to formalize and automate this.

There are some attributes of various behavior modes we should test for:

Continuous

• No sudden jumps
  All stock values should meet the definition of continuous, while flows can be adjusted with pulse functions, etc.

Smooth

• No sudden jumps in first derivative

Oscillating (vs. non-oscillating)

• The timeseries goes up and down

Periodic (vs. aperiodic)

• Oscillations have a consistent frequency

Stationary

• The envelope is constant

Divergent

• The envelope increases in size

Convergent

• The envelope decreases in size

We could also test for some specific behavior cases:

• Exponential Growth