Analog Quantum Simulation
Engineering a controllable quantum device whose native dynamics directly emulate a target Hamiltonian, without decomposing into gates.
Emulation, not computation
Analog quantum simulation builds a physical system, cold atoms, trapped ions, superconducting arrays, or Rydberg tweezers, whose intrinsic Hamiltonian is tuned to match the model of interest. The device then evolves under its own physics, and measurements read out the emulated dynamics. There is no gate decomposition; the hardware is the algorithm.
Leading platforms
- Ultracold atoms in optical lattices: emulate Hubbard and Bose-Hubbard models.
- Trapped ions: realize long-range Ising and spin models with tunable range.
- Rydberg-atom arrays: programmable geometries with strong van der Waals coupling.
- Superconducting resonator lattices: photonic and spin-boson models.
Strengths
Because the dynamics are native, analog simulators can maintain coherence over longer effective times and larger particle numbers than digital circuits on the same hardware generation. They have already reached regimes, hundreds of interacting spins, where classical simulation is difficult, giving early scientific results on quench dynamics and phase transitions.
Weaknesses
Analog simulators are special-purpose: each device emulates a limited family of Hamiltonians set by its physics. Errors are continuous and hard to correct, since standard quantum error correction assumes discrete gates. Calibration errors bias results, and verifying an analog result in a classically hard regime is itself a challenge. Precision is limited by control accuracy rather than a tunable error budget.
The near-term role
For models that match a platform's native interactions, spin lattices, Hubbard physics, gauge theories, analog simulation is the most capable near-term tool, delivering many-body science before fault-tolerant digital machines mature. The two paradigms are complementary rather than competing, and digital-analog hybrids seek the coherence of native evolution plus the programmability of interspersed gates. For long-horizon, general problems, digital simulation with error correction remains the endpoint.