"State of the Field" section in JOSS paper

[N-Wouda]

From the review in https://github.com/openjournals/joss-reviews/issues/5028:

State of the field: Do the authors describe how this software compares to other commonly-used packages?

And reviewer comment:

See the JOSS required item above to please add "State of the Field" to cover how this package compares to other packages. If none such exists (which I doubt it, since stochastic local search is well-known), please make the contribution explicit. If there are other tools, please cover mainstream ones, if any and mention where each would benefit the most (e.g., genetic algorithms for multi-point solutions vs. alns)

There's some references already written down in the review that might be useful. I'm thinking ten-or-so references at most (incl. some open source Python/other optimization packages), and text similar in size to the statement of need.

@leonlan can you have a go at this?

[leonlan]

Sure thing. Will work on this tomorrow!

[leonlan]

Packages that facilitate single-trajectory, single-objective optimization:

• https://github.com/jMetal/jMetalPy
• https://github.com/jmejia8/Metaheuristics.jl
• https://github.com/ac-tuwien/pymhlib
  • See https://github.com/ac-tuwien/pymhlib/blob/master/pymhlib/alns.py for alns
• https://github.com/nojhan/paradiseo
• https://joss.theoj.org/papers/10.21105/joss.02448

Other popular libraries but not directly easy to implement ALNS

• https://github.com/DEAP/deap
• https://github.com/ljvmiranda921/pyswarms
• https://github.com/anyoptimization/pymoo

Not a package but relevant

• https://github.com/alberto-santini/adaptive-large-neighbourhood-search

[leonlan]

I think the list above includes the most relevant packages.

In this branch, I slightly modified the text in the "statement of need" and added a single line of text to refer to other libraries. This is what many JOSS papers seem to do as well, and IMO it feels enough.

It feels a bit of a stretch to write a section on "State of the Field", because none of these packages are completely dedicated to single trajectory and single objective algorithms. But I think I can also make an attempt to write this.

I'll have a better look at this tomorrow, I need to fix my jetlag :-)