Convergence tracking

[quaquel]

This pull request entails an overhaul of how convergence tracking is being supported by the workbench. This closes #187.

In short:

1. ArchiveLogger now stores results in a tarbal
2. Additional helper functions for converting from a dataframe to a collection of platypus solutions
3. Additional helper function for merging results from multiple seeds using epsilon non dominated sort
4. All platypus convergence indicators are wrapped using a wrapper class to make them play nice with the workbench.
5. HyperVolume results in a DeprecationWarning being logged. It is better to calculate hypervolume in post. This also closes Docs: Expand on HyperVolume minimum and maximum parameters #149
6. modifications to various init files to ensure the correct functions and classes can be directly imported from ema_workbench.

this pull request needs to be merged before we can resolve issue #147

[coveralls]

Coverage decreased (-0.4%) to 81.139% when pulling 5f8274d on convergence_tracking into 3f60527 on master.

[EwoutH]

Looks like an major effort! I will dive into this tomorrow.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[EwoutH]

Two last nick picks, when those are resolved it's good to go!

[quaquel]

Two last nick picks, when those are resolved it's good to go!

A nice use case for the walrus operator indeed. I had completely. forgetting about it while writing the code.

[EwoutH]

I checked the Read the Docs and found a few missing empty lines in the docstring, fixed those in 5f8274d, and merged!

Congratulations on getting this major feature in!