New SciML/Optimization for fitting parameters of ODEs tutorial tutorial

[TorkelE]

A new version of this one: #708

Adds a tutorial of how to use DiffEqParamEstim and Optimization for parameter fitting. Those packages are still not properly adapted to symbolic indexing, so normal vectors have to be used. However, I still think we go with that (I can add an additional note box at the beginning if we want to highlight this).

I still think this is better than self-coding this stuff (easy-to-miss stuff like what happens if a solver fails) so think this is useful (PEtab is still probably the best alternative). Also, hopefully, we can use this as a template workflow that we think should work without indexing, and something DiffEqParamEstim + Optimization should work towards fixing.

[isaacsas]

What about fitting multiple data sets that should have the same parameters simultaneously? (i.e. if the experiment has been run multiple times) Do we show this anywhere?

[isaacsas]

Suggested change

	using OptimizationOptimisers # Required for the ADAM optimizer.
	using SciMLSensitivity # Required for `Optimization.AutoZygote()` automatic differentiation option.
	using OptimizationOptimisers # for the ADAM optimizer
	using SciMLSensitivity # for Optimization.AutoZygote() AD backend

[TorkelE]

I prefer the old version. Especially spelling out the AD part (As I imagine most people are unfamiliar with this, and especially just "AD backend" doesn't say much.

[TorkelE]

Also, I generally try to make my comments full sentences, not sure if there is some standard regarding this though.

[TorkelE]

What about fitting multiple data sets that should have the same parameters simultaneously? (i.e. if the experiment has been run multiple times) Do we show this anywhere?

I think I looked at demonstrating this as well, but it was just really messy so I decided not to do it (was a while ago, so don't really remember). There is a section mentioning that this is possible, and linking to where DiffEqParamEstim describes it though:

## Parameter fitting to multiple experiments
Say that we had measured our model for several different initial conditions, and would like to fit our model to all these measurements simultaneously. This can be done by first creating a [corresponding `EnsembleProblem`](@ref advanced_simulations_ensemble_problems). How to then create loss functions for these are described in more detail [here](https://docs.sciml.ai/DiffEqParamEstim/stable/tutorials/ensemble/).

We also show how to do this for PEtab.

[TorkelE]

Things updated and rebased on the latest v14 branch, thanks for the input!

[isaacsas]

Feel free to merge when you feel this is done and sufficiently revised.

[TorkelE]

Thanks for the input :)

I will have a second read-through tonight just to be sure, and will merge afterwards.