R2 optimality measure

[MaxenceGollier]

Adapted R2 optimality measure

[dpo]

@geoffroyleconte @MohamedLaghdafHABIBOULLAH Could you please look into this? I think we should do the same in all solvers.

[codecov]

Codecov Report

Patch coverage is 100.00% of modified lines.

Files Changed	Coverage
src/R2_alg.jl	100.00%

📢 Thoughts on this report? Let us know!.

[github-actions]

Here are the
demos-results

[geoffroyleconte]

I find that for my experiments, it is not very satisfying, especially for TR (maybe we would have to adjust the tolerances of the subsolver). This makes the stopping criteria more difficult to reach (but $\xi_1$ is smaller in the end). Also, there is the problem of the "noredTR" variant in TRDH, i'm not sure how we should deal with $\xi$ to terminate the algorithm in this case.

[dpo]

Let's not worry too much about "noredTR" because we don't have a convergence proof for it at this point.

[dpo]

Thank you!