qaoa demo

[Zhaoyilunnn]

1. Implemented estimator on simulator.
2. Add a demo in unit test.

Now the APIs in algorithms module are in very early stage.
There may be significant changes in the future, depending on

1. Whether we need to integrate other DL platforms like tf and torch.
2. Will we have different communication protocol when submitting VQA jobs.

[chensgit169]

In simulation, $\langle\psi|$ is straightforward to be obtained by .conj(). However, in experimental circuits, it needs to be eplemeted manually by inverse the circuit that builds $|\psi\rangle$. Thus it seems there is still a method .inverse() to be implemented in ansatz circuit.

[Zhaoyilunnn]

In simulation, ⟨ψ| is straightforward to be obtained by .conj(). However, in experimental circuits, it needs to be eplemeted manually by inverse the circuit that builds |ψ⟩. Thus it seems there is still a method .inverse() to be implemented in ansatz circuit.

Thanks! I am still seeking the best implementation for calculating expectations on real devices. Previously shangshu has an implementation, my understanding is that we do not need to obtain $\langle\psi|$ (and maybe we are also not able to do so?) in experiment. Let me know if I misunderstood your comment...

Anyway I am adding a TODO for experimental implementation, currently I did some tricky conversion and reuse the Task.submit