Drop python 3.8 support, add python 3.12 tests, and pin qiskit<2

[t-imamichi]

Summary

Addresses #638
Fixes #657

qiskit-machine-learning also dropped the support of Python 3.8 already.
qiskit-community/qiskit-machine-learning#826

This adds tests of python 3.12 for Linux and Windows because the latest CPLEX provides wheels for the platforms.
Note that CPLEX does not provide a wheel for Python 3.12 on macOS.

• https://pypi.org/project/cplex/22.1.2.0/#files (only Windows and Linux)

This also simplifies gurobipy constraints because gurobipy provides arm64 wheels.

• https://pypi.org/project/gurobipy/12.0.1/#files

Details and comments

[coveralls]

Pull Request Test Coverage Report for Build 14508752889

Details

• 1 of 1 (100.0%) changed or added relevant line in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.05%) to 92.104%

---

Totals
Change from base Build 13105033536:	0.05%
Covered Lines:	5681
Relevant Lines:	6168

---

💛 - Coveralls

[TolisChal]

Thanks @t-imamichi! This PR looks good to me.

[t-imamichi]

Added a release note

[Dniskk]

Looks good!

[TolisChal]

LGTM!

[woodsp-ibm]

Seems fine to me. The tox.ini py313 test env, I guess it doesn't hurt, just seems a bit counter to the max of 3.12.

Of course the Branch Rules will need updating to match these changes.

[woodsp-ibm]

@t-imamichi While I approved it above a while back I did not merge this or anything as I guess I was expecting someone else to do that. Of course as I noted the branch rules would need changing to match what this does. Now since then Qiskit 2.0 has been released and the nightly Actions here fail due to that, as well as the fact that under 3.8, which this PR removes, its trying now to build Qiskit Aer as the latest no longer has 3.8 support and released wheels. If this were re-run then it would fail due to Qiskit 2.0 - so the merge Action would too. In other repos the Qiskit version has been pinned to add <2.0 so it will not install that version until such time if/when V2 primitives etc are supported by the application. I was wondering if it would be a good idea to add that pin to this PR such that hopefully, as long as there is no other issue, CI would pass again overall.

[t-imamichi]

Thank you, Steve. I forgot to merge it when approved. I add a pin.

[t-imamichi]

I updated CI constraints as well.

[ElePT]

LGTM!

[woodsp-ibm]

Thanks for updating the PR Imamichi-san, and the branch rules. Looks good.

[t-imamichi]

Thank you for reviewing this, Steve and Elena!