Variational Quantum Eigensolver (VQE)

A hybrid quantum-classical method that finds ground-state energies by minimizing the expectation of a Hamiltonian over a parameterized circuit.

The variational principle

For any normalized state |psi>, the expectation is at least the ground-state energy E_0, with equality only for the ground state. VQE (Peruzzo et al., 2014) exploits this: parameterize a trial state |psi(theta)> with a quantum circuit, measure its energy, and use a classical optimizer to adjust theta downward toward E_0.

The hybrid loop

• Prepare |psi(theta)> on the quantum device with a parameterized ansatz circuit.
• Measure the expectation of each Pauli term in H = sum_l c_l P_l by repeated sampling.
• Combine to estimate the total energy E(theta) = sum_l c_l .
• Feed E(theta) to a classical optimizer, which proposes new theta.
• Repeat until the energy converges.

Why it suits near-term hardware

VQE circuits are shallow and the classical optimizer absorbs some noise, so it runs on noisy devices that cannot support deep phase-estimation circuits. It sidesteps the long coherent evolution that fault-tolerant algorithms require.

Measurement cost

The catch is sampling: estimating each to precision epsilon needs O(1/epsilon^2) shots, and molecules have many terms. Grouping commuting Paulis, classical shadows, and importance sampling reduce this, but measurement overhead is a central VQE bottleneck alongside optimizer difficulty.

Known limitations

VQE can stall in barren plateaus, regions where gradients vanish exponentially in the number of qubits, making optimization intractable for expressive random ansatze. Local cost functions, structured ansatze, and careful initialization mitigate this. VQE also gives no rigorous accuracy guarantee; it returns an upper bound whose quality depends on the ansatz. It remains the leading approach to quantum chemistry on pre-fault-tolerant machines, and a proving ground for hardware.