Fix broadcasting bug for rolling horizon with linearised UC case

[Irieo]

Makes optimization case raised in #1454 possible

Changes proposed here

A small fix for an xarray broadcasting error in the rolling-horizon & linearised unit commitment case.

When running optimize_with_rolling_horizon(linearized_unit_commitment=True, horizon=...) with both committable and non-committable generators, the ramp-limit constraints for committable units tried to combine two arrays with mismatched name dimensions: rhs was built only over committable generators, and the initial status_start used for the previous rolling window still carried all generators (committable + non-committable).

The fix is straightforward: align status_start to the committable index before using it in the RHS for the ramp-up constraint.

Also add unit tests for this; these combination of optimization settings was not covered.

Checklist

• Code changes are sufficiently documented; i.e. new functions contain docstrings and further explanations may be given in docs.
• Unit tests for new features were added (if applicable).
• A note for the release notes docs/release-notes.md of the upcoming release is included.
• I consent to the release of this PR's code under the MIT license.