Special patterns of constraint to penalty converters

[knamba-jp]

Summary

Resolves #40

TODO

• Implementation of the converter
• Intuitive converter name
• Docstring
• Unit test

Details and comments

Owned by Amano-san and me

A Tutorial on Formulating and Using QUBO Models introduces examples of the special patterns in page 10 as follows.

[CLAassistant]

All committers have signed the CLA.

[t-imamichi]

We need a more intuitive name than SpecialConstraintToPenalty because there might be other types of inequality to equality converters in the future.

[t-imamichi]

Could you apply black code formatter by black -l 100 qiskit_optimization.
We recently apply black to the entire code base. See #117 #118 for details.

[ibmamnt]

We will run black command as suggested. As for the name of the class.
@t-imamichi How about this name ?

KnownConstraintConverter

[t-imamichi]

Thanks. The code format looks good. The naming is a bit difficult. What is the best name for this converter...? Maybe we can leave it later.

[t-imamichi]

LinearInequalityToPenalty may work.

FYI, you don't have to implement patterns x+y=1 and x=y because LinearEqualityToPenalty can handle them.

[t-imamichi]

You can simplify it by using PauliSumOp.from_list.
https://qiskit.org/documentation/stubs/qiskit.opflow.primitive_ops.PauliSumOp.html?highlight=paulisumop#qiskit.opflow.primitive_ops.PauliSumOp.from_list

[knamba-jp]

We discussed and concluded it should be done in another PR.

[knamba-jp]

Since some errors can't be fixed, we'll open another PR.