[SCIP] new handling of nonlinear constraints in SCIP/master

[Stefan Vigerske]

Dear early adopters,

a rewrite of the handling of nonlinearities (as in NLP/MINLP) in SCIP 
and the addition of new related features has been merged into the 
master-branch of SCIP, and thus is available at 
https://github.com/scipopt/scip. Note, that the master branch is the 
main development branch of SCIP.

The new code can result in significant changes in performance (in any 
direction). Further, many parameters were added or removed and the API 
to work with expressions, nonlinear constraints, and interfaces to NLP 
solvers has changed a lot. The CHANGELOG lists added and removed 
features, API functions, and parameters.

For some overview on the motivation, new design, and features added, I 
refer to slides from some previous presentations:
- https://www.zib.de/vigerske/consexpr_ifors2021.pdf  (2021)
- https://www.scipopt.org/workshop2020/slides/minlp.pdf (2020)
- https://www.zib.de/vigerske/consexpr_wcgo.pdf  (2019)

Enjoy,
Stefan (for Antonia, Benny, Fabian, Felipe, Ksenia, and me)