Long-term STEP improvements

[DavidSnyder-TRI]

Longer-term TODOs:

• Script to automatically find a good set of parameters (lambda_value, major_axis_length) for STEP(n_max, alpha)
• Policy and risk accumulation visualizations (very easy with the older, inefficient policy representation)
• Estimates of memory / runtime built into the user interface, as well as restart if synthesis is buggy (e.g., for bad parameters -- see 1, above)

[DavidSnyder-TRI]

(1) and (2) addressed in PR #26, auto-restarting the policy synthesis remains a TODO.