Non-variational ADAPT Algorithm For State Preparation

The ADAPT-VQE algorithm grows the ansatz iteratively by selecting the operator with the largest measured energy gradient.

A challenge with any VQE is the variational optimization, which may pose a barrier to scalability.

Here, we explore a variant of ADAPT that estimates the variational parameters based on the measurements made in the previous part for growing the ansatz and hence replace the VQE optimization subroutine.

We compare the rate of energy lowering, circuit depth, and measurement cost for molecular models of the original ADAPT-VQE algorithm and its non-variational counterpart.