Algorithm Zoo · Simulation
Trotterized Hamiltonian Simulation
Also known as: Lie-Trotter-Suzuki, Product-formula simulation
First described: Seth Lloyd (universal); Suzuki (higher-order), 1996
The problem
Implement e^{-iHt} for a local Hamiltonian H.
Approximate the time-evolution operator by Trotterizing the sum H = Σ Hk into a sequence of e^{-iHk Δt} pieces. First-order error is O(t²/r) for r steps; Suzuki's recursion gives p-th order error O((t/r)^{p+1}).
Best classical
Exponential in number of particles for fermionic / many-body systems
Quantum complexity
Õ(L·t² /ε) for first-order; Õ(L·t·(t/ε)^{o(1)}) for high-order
Our verdict
The most defensible quantum advantage application. Tensor networks have eaten more 'quantum supremacy' claims than the field would like to admit, but truly hard chemistry and condensed-matter problems remain beyond classical reach. Trotter is the workhorse — qubitization is the polished tool.
When to use it
- Quantum chemistry — pre-cursor to VQE for ground-state finding, full QPE for excited states once FTQC arrives.
- Lattice models (Hubbard, Heisenberg) where local Hamiltonians make Trotter natural.
- Time evolution under known Hamiltonians for dynamical-property estimation.
When not to use it
- Highly non-local Hamiltonians where the Trotter step has unfavorable scaling.
- When you can replace Trotter with qubitization / LCU for tighter bounds.
Classical baseline
Tensor networks (MPS, PEPS) for 1D and weakly entangled 2D systems are extraordinarily strong — many systems formerly cited as 'quantum supremacy candidates' have been classically simulated since.
Hardware cost
Per Trotter step: ≈L two-qubit gates for an L-local Hamiltonian. Simulating a 50-site Hubbard model to chemical accuracy is estimated to need depth ~10^5 — feasible only post-FTQC.
Key papers
- Universal quantum simulators ↗
Lloyd · 1996 · Science
- Theory of Trotter Error with Commutator Scaling ↗
Childs, Su, Tran, Wiebe, Zhu · 2021 · Phys. Rev. X
Deep-dive tutorials
Last verified: 2026-05-24