Noise-adaptive initial qubit layout pass

[prakashmurali]

This pass uses backend calibration data to find a good initial qubit layout which will improve program success rate. The pass is based on the greedy heuristic from
Prakash Murali, Jonathan M. Baker, Ali Javadi-Abhari, Frederic T. Chong, Margaret R. Martonosi. Noise-Adaptive Compiler Mappings for Noisy Intermediate-Scale Quantum Computers, ASPLOS 2019
(https://arxiv.org/abs/1901.11054)

Summary

The pass uses backend calibration data to determine a good set of physical qubits for the program dag. It selects physical qubits based on CNOT and readout errors. The pass improves the program's success rate (fraction of runs that produce the correct answer) on real hardware evaluations as demonstrated in arXiv:1901.11054.

Details and comments

This pass implements the noise-adaptive initial layout technique, it does not perform noise-adaptive swapping.

✅ I have added the tests to cover my changes.
✅ I have updated the documentation accordingly.
✅ I have read the CONTRIBUTING document.

[CLAassistant]

All committers have signed the CLA.

[CLAassistant]

Thank you for your submission, we really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[ajavadia]

Thanks @prakashmurali! The code is nicely written and very easy to follow.
I left a few minor comments inline, but I think we can merge this soon.

I think one thing to perhaps remember/document here is that this pass considers all 2-qubit gates in a given circuit, but at the hardware level it currently only considers CX gates (i.e. a virtual CZ is mapped to the best physical CX). In the future we should update this to consider other physical gates (let's say for a hardware that has iSWAP or whatever).

[prakashmurali]

The Travis build failed for test.python.basicaer.test_qasm_simulator.TestBasicAerQasmSimulator.test_qasm_simulator
It is unrelated to the changes in this PR.

[ajavadia]

Ok great, I'll merge this!

We'll run some benchmarks and my guess is this is going to perform better than the current layout, in which case we'll make this the default for devices.

[prakashmurali]

Thank you very much for the quick review and responses!