Skip to content
Quantum.world

The Signal

Gumaro Rendón’s Work Enables Nearly Linear-Cost Estimation of Observables After Hamiltonian Evolution

A method combining Trotter product formulas and interpolation estimates observables after Hamiltonian evolution with nearly linear cost scaling in simulation time. It uses whole-number Trotter steps to reconstruct fractional operations, avoiding extra block-encoding qubits.

Gumaro Rendón’s Work Enables Nearly Linear-Cost Estimation of Observables After Hamiltonian Evolution

Observable estimation has nearly linear cost scaling in simulation time

A Trotter-based method estimates observables after Hamiltonian evolution with cost scaling of Õ(T1+o(1) log2(1/ε)/ϵ), while its maximum depth scales as Õ(T1+o(1) log(1/ε)/ϵ). Here, ε is the algorithmic error in the target evolution, ϵ is the statistical error in estimating an observable, and Õ suppresses factors that are doubly logarithmic in 1/ε.

Source quantum-journal.org

The paper, “All you need is Trotter,” was published in Quantum on November 6, 2026, by Gumaro Rendón. The work enables estimation of observables after an evolution written as eiHT without extra block-encoding qubits.

Quantum computers can simulate the evolution of a system and measure a property such as an energy or correlation. Trotter’s method breaks an evolution into small time steps made from simpler pieces. It is described in the paper’s popular summary as a simple, hardware-friendly approach, but reducing its error requires smaller steps and deeper circuits. More sophisticated techniques can reach high precision at lower cost, but require extra qubits and circuitry.

Whole-number steps evaluate fractional quantum operations

The method combines product formulas with stable Chebyshev interpolation. The interpolation can call for a fractional number of Trotter steps, which a circuit cannot run directly. The approach gets around that requirement by using the circuit to run whole-number step counts and then reconstructing the needed fractional operation from the results.

For a fractional operation of 7.42 steps, for example, the quantum computer runs nearby integer counts such as 5, 6, 7, 8 and 9 and measures each result. A classical computer combines the measurements using cardinal sine interpolation. The paper compares the reconstruction’s role to the way digital audio can rebuild a smooth sound wave from discrete samples.

A Gaussian window keeps the number of samples required small, and the reconstruction is described as accurate and stable against measurement noise. The extrapolation uses carefully spaced points to estimate the result at zero step size, where the Trotter error disappears. This avoids running a fractional step on the quantum hardware.

Similar interpolation methods have superlinear cost scaling

Similar interpolation methods have a superlinear cost scaling of 5pT1+1/p, where p is the product-formula order. The work uses an optimized order p that varies with simulation time T.

The paper describes its approach as an improvement in required qubits over Quantum Signal Processing methods while retaining scaling with polylog(1/ε). Its use of cardinal sine interpolation gives a slightly better log2(1/ε) factor in the cost than the log3(1/ε) factor in other methods using Chebyshev interpolation and product formulas.

The work provides an alternative for fractional queries

The paper also presents cardinal sine interpolation as an alternative way to calculate observables or amplitudes after fractional-query evolution, with cost O(T log(1/ε)). Its popular summary describes the method as using Trotter circuits and two extra qubits. The approach is presented as suited to early fault-tolerant quantum computers, where qubits and additional circuitry may be scarce.

Source

  1. Quantum

Related stories

Explore

More articles