Fix NLOpt optimizers trying to iterate NoneType

[woodsp-ibm]

Bounds were assumed to be a an array of tuples. No bound being (None, None) tuple for each param. Passing None as the bounds array (default when not explicitly set) always iterated through bounds to set thing up and failed trying to iterate NoneType. This was reported in external Slack workspace https://qiskit.slack.com/archives/C7SJ0PJ5A/p1602918563027800 (if you cannot access the link and want to there is a link here to join the community https://github.com/Qiskit/qiskit-aqua#contribution-guidelines ). Here is a summary though

Hi: Has someone tried global optimizers from Aqua? I followed the instruction from https://qiskit.org/documentation/apidoc/qiskit.aqua.components.optimizers.html#global-optimizers and installed NLopt but got this error during running: op = optimizer.optimize(num_vars=num_params, objective_function=init_cost0, initial_point=params)
File "/p/home/liux3790/anaconda3/envs/name_of_my_env/lib/python3.8/site-packages/qiskit/aqua/components/optimizers/nlopts/nloptimizer.py", line 96, in optimize
return self._minimize(self._optimizer_names[self.get_nlopt_optimizer()],
File "/p/home/liux3790/anaconda3/envs/name_of_my_env/lib/python3.8/site-packages/qiskit/aqua/components/optimizers/nlopts/nloptimizer.py", line 124, in _minimize
low = [(l if l is not None else -threshold) for (l, u) in variable_bounds]
TypeError: 'NoneType' object is not iterable.