Parallel circuit implementation of variational quantum algorithms

Abstract

We present a framework to split quantum circuits of variational quantum algorithms (VQAs) to allow for parallel training and execution to solve problems larger than the number of available qubits in a quantum device. We apply this method to combinatorial optimization problems, where inherent structures can be identified, and show how to implement these parallelized quantum circuits. We show how to formulate an objective function for the classical optimizer to guide the optimization towards meaningful solutions. We test our framework by creating a parallelized version of the Quantum Approximate Optimization Algorithm and a variational version of quantum annealing and explain how our framework applies to other quantum optimization algorithms. We provide results obtained both from simulation and experiments on real hardware. Our results show that the information lost by splitting the quantum circuits can be partially recovered by optimizing a global objective function evaluated with the separate circuit samples.