Add best known solution import for the VRP

[swagner]

An import functionality should be added for the best known VRP-solution.

[svonolfe]

Add best known solution import for the VRP (TSPLib format) in r4619, Improved best known solution wiring in r4620, fixed backwards compatibility in r4625, fixed null reference exception in r4626

[svonolfe]

To test the functionality: Import a VRP instance and a best known solution (TSPLib-Format) from ​http://www.branchandcut.org/VRP/data/

[abeham]

As discussed with svonolfe in the review, a few changes and bugs still need to be fixed. I'm reassigning this ticket so that it appears as waiting for completion.

[svonolfe]

Implemented review comments in r4847

[abeham]

I've identified a few more issues:

• The BestVRPSolutionAnalyzer should update BestKnownQuality and BestKnownSolution in case they're not available.
  • Selecting BestKnownSolution should show a view of the solution, maybe together with an export button "Export as optimal solution"
• In the import dialog when a new VRP instance is loaded, the field for the optimal solution should be emptied
• The message box that says "Invalid solution" should be a bit more descriptive saying e.g. "The optimal solution does not seem to correspond with the problem data." in case it treats a different number of customers or "The optimal solution file is not in the correct format, please visit [link to our wiki page] for a description of the file format."
  • A description of the file format should be on the wiki page (an example should be enough)

[svonolfe]

Implemented review comments in r4851

[abeham]

Using the solution object as data type for BestKnownSolution is not optimal since it displays the same functionality as the visualization tab in the problem view. Changing the datatype back to a representation object and implementing a visual representation (a textual form of the routes) for IVRPEncoding would be generally useful. Such a view could then later be extended to include an "Export as optimal" button.

[svonolfe]

Partyl reverted r4851 in r4852, implemented review changes in r4856

[svonolfe]

I display the genotype in its textual form directly now: in the case of the Potvin encoding the routes, in the case of the Alba encoding the underlying Permutation.

[abeham]

Excellent, I think this is ready for release.