Research questions to function mapping

[CIGbalance]

This is a first draft for my task in #4, relating to the research questions that can be answered by comparing across different functions. We agreed that I would first add something to the webpage (which is this PR, and then we could potentially include similar groupings in cocopp).

Before converting this to a full PR, I would like to discuss the following:

• Do the visualisations make sense? I figured for some points (like separability), it is nice to see. But I am currently only showing 2D, not sure if that is too simplistic
• Do the headers make sense? They are mostly just what is already in the documentation, but not entirely. I think we should also add more details to a lot of them, but that is secondary for now.

[nikohansen]

I think it's currently difficult to understand that there are two semantic approaches and (only) the second compares two cases (or investigates a specific change).

The Rosenbrock function should not be characterized as separable or an example of exploitation of separability. One could say that we test whether a dependency on the given coordinate system is exploited. The 2-D case isn't likely to reflect the question in this case too.

[CIGbalance]

I think it's currently difficult to understand that there are two semantic approaches and (only) the second compares two cases (or investigates a specific change).

That is a very good point. Just to clarify, you mean the difference between (1) observing optimal convergence by looking at the sphere function only, vs (2) comparing the behaviour across 2 similar functions, right? I'll think about how to make this clearer. I am in generally not entirely happy with the presentation, but haven't thought of a good alternative yet.

The Rosenbrock function should not be characterized as separable or an example of exploitation of separability. One could say that we test whether a dependency on the given coordinate system is exploited. The 2-D case isn't likely to reflect the question in this case too.

Good point, I'll clarify 👍 (at some point, when I return to this 😇 )