Trotter–Suzuki Product Formulas
Splitting e^{-i(A+B)t} and bounding the error.
Product formulas are the oldest way to simulate quantum dynamics and still the most practical: no ancillas, no oracles, just exponentials of the terms you already have. This lesson builds Suzuki's formulas of every even order from one symmetry argument and proves the commutator-scaling bounds that let them compete with asymptotically optimal algorithms.
What you’ll learn
- Define the order of a product formula and prove that time-symmetric formulas have even order.
- Derive Suzuki's recursion and its coefficient , and explain why negative time steps are unavoidable beyond order two.
- State the commutator-scaling theorem of Childs, Su, Tran, Wiebe & Zhu and sketch its proof.
- Derive the gate count for lattice Hamiltonians and compare product formulas with post-Trotter methods.
Product formulas and their order
The Lie–Trotter formula has order 1. The symmetric Strang formula
has order 2. You could verify that with the Baker–Campbell–Hausdorff formula, but there is a slicker argument that works at every order.
For small , BCH writes any product of exponentials as one exponential of a power series. First-order consistency ( for each ) fixes the linear term.
Substitute ; invert by negating the exponent.
Near the identity the logarithm is unique, so the exponents agree term by term. The first surviving error is at an odd power : the order is even.
Which of these is a second-order approximation of for non-commuting , ?
Show the answer
B.
It is first-order consistent and palindromic, so and the symmetry theorem forces order 2. Which operator gets split does not matter.
Suzuki's fractal recursion
Suzuki (1990, 1991) raised the order two at a time with a palindrome of five copies of the previous formula:
By the symmetry theorem the leading error of the symmetric formula sits at the odd power .
The leading exponents are all multiples of the same , so they commute; the error terms add at leading order, since their commutators with only enter at order .
Consistency is automatic; the second equation kills the error.
Take real -th roots of .
The arrangement is palindromic, so is again symmetric, and the term vanishes too. Hence has order .
For the fourth-order formula (), compute the time coefficient of the middle stage.
Show the answer
-0.66
, so . The middle stage runs time backwards. (Numerically, halving shrinks the error of by a factor , confirming order 4.)
The price of the recursion is exponential in : contains copies of , about exponentials per step. That factor is why practitioners rarely go beyond order 4 or 6.
Commutator scaling
Early analyses (Lloyd 1996; Berry, Ahokas, Cleve & Sanders 2007) bounded the per-step error by powers of . That is wasteful: in a lattice most pairs of terms commute, and commuting terms contribute no error at all. The right currency is nested commutators.
Variation of parameters (Duhamel). It suffices to bound .
Flatten into exponentials . Differentiating puts where its exponential sits; moving it to the front conjugates it by the product of the exponentials to its left.
Expand each conjugation in nested commutators , with an exact integral remainder.
The order conditions are exactly the statement that the polynomial part cancels up to degree . Only remainders survive, and each is bounded by norms of nested commutators of terms.
Integrate. For this reproduces the Excited State bound with explicit constant: .
Lattices. For a nearest-neighbor chain of sites with , a nested commutator is nonzero only if its terms form a connected cluster, so for fixed , versus for the norm-based bound. Hence
almost linear in the space-time volume . Haah, Hastings, Kothari & Low (2018) reached with a different decomposition and showed this is optimal up to polylogarithmic factors.
A nearest-neighbor chain of qubits is simulated for time at fixed precision with a fourth-order formula. Using commutator scaling, how does the total gate count scale?
Show the answer
B.
Gates with and : .
Practice and frontier
Asymptotically, product formulas lose to post-Trotter methods: their cost grows like rather than . In practice they often win. They need no ancillas, their constants are small, commutator scaling exploits locality, and the actual error is often far below the bounds. Childs, Maslov, Nam, Ross & Su (2018) compiled concrete Heisenberg-chain simulations and found product formulas, especially with empirically estimated error, needed fewer gates than the asymptotically superior algorithms at the sizes of interest.
| Method | Steps or queries for time , error | Key feature |
|---|---|---|
| Lie–Trotter () | simplest | |
| Suzuki, order | exponentials per step | |
| qDRIFT (Campbell 2019) | exponentials, | independent of the number of terms |
| Truncated Taylor series (LCU) | block encodings, ancillas | |
| QSP / qubitization | optimal query complexity |
Key takeaways
- A -th order product formula errs by per step, so steps suffice (Childs, Su, Tran, Wiebe & Zhu 2021).
- Time-symmetric formulas have even order; Suzuki's palindrome with raises the order by two at five times the cost.
- Beyond order two, negative time steps are unavoidable (Sheng–Suzuki).
- Commutator scaling makes lattice simulation nearly linear in the space-time volume, and in practice product formulas often beat asymptotically optimal methods.
Check yourself
In the interactive lesson these are scored with your confidence, and what you miss comes back for review.
A product formula satisfies and . What can you conclude?
- It is first order
- It is exact
- It is at least second order
- It is fourth order
Show the answer
C. It is at least second order
Time-reversal symmetry kills every even power in , including , so the first error is at least . It need not be fourth order: Strang's formula is symmetric and has a nonzero term.
Why does Suzuki's fourth-order formula contain a backward time step?
- Every product formula of order must contain a negative coefficient; here
- To undo decoherence accumulated in the forward steps
- Because must be made traceless first
- It does not; all five stages run forward in time
Show the answer
A. Every product formula of order must contain a negative coefficient; here
Cancelling the error requires , which forces . More generally, the Sheng–Suzuki theorem says no all-positive splitting exceeds order two.
Why are product formulas beyond order 4–6 rarely used?
- Higher orders are not more accurate for small steps
- Negative time steps make them non-unitary
- Commutator scaling fails beyond order 6
- The number of stages grows like , which outweighs the better scaling at realistic precisions
Show the answer
D. The number of stages grows like , which outweighs the better scaling at realistic precisions
Each order increase multiplies the exponentials per step by 5. The exponent of only improves from to , which pays off only at extreme precision or very long times.
For a nearest-neighbor chain of sites with bounded terms, how does scale with at fixed order ?
Show the answer
B.
A nested commutator of terms vanishes unless the terms form a connected cluster. There are starting positions and clusters of bounded size at each, so the sum is .
Further reading
- H. F. Trotter, "On the product of semi-groups of operators", Proceedings of the American Mathematical Society 10, 545 (1959).
- M. Suzuki, "General theory of fractal path integrals with applications to many-body theories and statistical physics", Journal of Mathematical Physics 32, 400 (1991).
- D. W. Berry, G. Ahokas, R. Cleve & B. C. Sanders, "Efficient quantum algorithms for simulating sparse Hamiltonians", Communications in Mathematical Physics 270, 359 (2007).
- A. M. Childs, Y. Su, M. C. Tran, N. Wiebe & S. Zhu, "Theory of Trotter error with commutator scaling", Physical Review X 11, 011020 (2021).
- A. M. Childs, D. Maslov, Y. Nam, N. J. Ross & Y. Su, "Toward the first quantum simulation with quantum speedup", PNAS 115, 9456 (2018).