Add or tweak a converter to deal with trivial variables and constraints

What is the expected enhancement?

QuadraticProgram may include trivial variables and constraints. For example, #227 includes a trivial variable (lower bound and upper bound are equal).

Here is an example of a trivial constraint.

from qiskit_optimization.problems import QuadraticProgram

qp = QuadraticProgram()
qp.linear_constraint(sense='==', rhs=1)  # constraint: 0 == 1

It is good to eliminate such trivial cases in some existing converters or a new converter.

Fixes #227

[wpoxon-qci]

@wpoxon-qci I made another bug fix of int2bin #230. If you are OK, I can include your suggestion as part of #230. If you want to make your own PR, it’s also alright. What do you think?

@t-imamichi - if you are already working on a PR for another fix in the int2bin code, please go ahead and include the fix for this issue in your other PR. Thank you!

[wpoxon-qci]

@t-imamichi and @atulanand206 - Thank you again for your quick response and for investigating this issue and working on a fix!

@atulanand206 - are you thinking about a fix that looks something like this?

--- .pyenv/versions/3.8.6/lib/python3.8/site-packages/qiskit_optimization/converters/integer_to_binary.py.orig	2021-08-18 19:41:11.327849767 +0000
+++ .pyenv/versions/3.8.6/lib/python3.8/site-packages/qiskit_optimization/converters/integer_to_binary.py	2021-08-18 19:49:45.889513140 +0000
@@ -101,7 +107,10 @@
         self, name: str, lowerbound: float, upperbound: float
     ) -> List[Tuple[str, int]]:
         var_range = upperbound - lowerbound
-        power = int(np.log2(var_range))
+        if var_range:
+            power = int(np.log2(var_range))
+        else:
+            power = 0
         bounded_coef = var_range - (2 ** power - 1)
 
         coeffs = [2 ** i for i in range(power)] + [bounded_coef]