[vqe][scipy_minimizer] Add argument to input x0 value

[gyu-don]

Now, scipy optimizer wrapper set randomized parameter as x0 value.
However, user may know better initial value.

[gyu-don]

In current implementation of blueqat.vqe, when minimizer is called, a number of parameters is passed to minimizer.
This behaviour is reasonable if user doesn't know the detail of the ansatz.
However, if user knows the ansatz and wants to adjust parameters, this behaviour is inconvenient.

There are some ad-hoc workarounds.
Ignoring n_params is one of an easy workaround.

class ScipyMinimizer:
    def __init__(self, **kwargs):
        self.minimizer_options = kwargs
        self.params = [random.random() for _ in range(n_params)]
    def set_params(self, params):
        self.params = params
    def __call__(self, objective, n_params):
        assert n_params == len(self.params)
        reslt = scipy.optimize.minimize(objective, self.params, **self.minimizer_options)
        return result.x

But I have no idea for permanent measures for this issue.