QOC — Quantum Optimal Control

What is quantum optimal control?

Quantum optimal control asks how to shape the time-dependent fields that drive a quantum system — laser, microwave, radio-frequency or voltage pulses — so that it performs a task as well as possible: prepare a state, implement a gate, or sense a signal, within limits on time, power and bandwidth, and in the presence of noise.

The field grew out of nuclear magnetic resonance and the laser control of chemical reactions in the 1980s and 1990s. Today it is a core tool of quantum technologies, from superconducting and trapped-ion processors to atomic clocks, quantum simulators and sensors.

See how the field is organized

$$H(t) = H_0 + \sum_k u_k(t)\, H_k$$ $$\max_{u}\;\; \mathcal{F}\big[U(T)\big] \;-\; \lambda \sum_k \int_0^T u_k(t)^2\, dt$$

The controls $u_k(t)$ enter the Hamiltonian alongside the drift $H_0$. An optimizer adjusts them — here piecewise constant, as in GRAPE — to maximize the fidelity $\mathcal{F}$ of the final evolution $U(T)$ with respect to a target, while penalizing pulse power.

Optimal control and the circuit model

Quantum algorithms are usually written in the circuit model: a sequence of gates from a small, fixed set, each carried out by a calibrated pulse. Quantum optimal control works one level lower. It designs the pulses themselves — for one gate, for a block of gates, or for a whole task such as preparing a state or sensing a field. The two views are complementary: optimal control produces the gates the circuit model takes for granted, and pulse-level compilation lets a program use either where it pays off.

Designing pulses

Strengths

Uses the full dynamics of the hardware, so operations can be shorter and closer to the speed limit.
Can replace a whole block of gates with one pulse, shortening programs and their exposure to noise.
Handles device-specific effects — leakage, crosstalk, distortion — directly in the design.
Targets any task, not only unitary gates: state preparation, sensing, dissipative and open-system control.

Limitations

Needs an accurate model of the device, or costly optimization on the hardware.
Optimized pulses are specific to one device and harder to benchmark and verify.
Numerical design scales with the Hilbert-space dimension, so it is limited to a few qubits or to blocks.
Its errors are less regular, which complicates error-correction analyses.

Composing gates

Strengths

A hardware-independent abstraction: the same program runs on different devices.
A small gate set is calibrated once and benchmarked, for example with randomized benchmarking.
Error correction and fault tolerance are built on gates with simple error models.
Scales by composition: large circuits are built from verified parts.

Limitations

Decompositions into native gates can be much longer than the operation requires.
Fixed gate times and idle periods spend coherence that near-term devices cannot spare.
The abstraction hides hardware features that a pulse could exploit.

Where the two meet: gate synthesis, pulse-level compilation and variational algorithms as control.

27 years of quantum control on arXiv

7,460 quantum control papers on arXiv since 2000

Papers in 2025: 706
2026 so far: 762
Share of all quant-ph papers (2025): 4.2%

Counted automatically from arXiv every day: preprints in quant-ph (and quantum papers in neighbouring categories) whose title or abstract uses quantum-control vocabulary. Hand-checked samples show the counts are approximate but consistent over time, so the trend is reliable. How papers are counted and checked.

QOC papers per year by date of first arXiv submission

Show the numbers

Year	Papers	1 Jan – 10/19	Share of quant-ph
2000	25	23	1.3%
2001	44	36	1.9%
2002	50	37	1.9%
2003	80	65	2.6%
2004	87	62	2.8%
2005	92	77	2.6%
2006	92	69	2.5%
2007	108	90	2.8%
2008	130	108	3.4%
2009	148	117	3.4%
2010	168	131	3.3%
2011	200	162	3.4%
2012	239	191	3.7%
2013	235	206	3.8%
2014	294	240	4.4%
2015	286	234	4.3%
2016	299	239	4.2%
2017	307	252	4.1%
2018	321	261	4.1%
2019	330	263	3.8%
2020	413	330	4.3%
2021	414	312	3.9%
2022	501	384	4.5%
2023	523	403	4.2%
2024	606	479	4.3%
2025	706	520	4.2%
2026 (to date)	762	762	4.4%

Everything about the field, in one place

Field map

An interactive map of the field: every area and method, how they connect, and an article per topic with the essential background and key papers.

Paper tracker

New quantum control papers from arXiv, collected automatically every day and classified into the topics of the field map.

Research groups

A world map of the groups working on quantum optimal control — who they are, where they are, and what they work on.

Open problems

A curated list of specific, well-posed open questions in quantum optimal control — with context, partial results and references.

Resources

Everything else you need: software packages, reviews, tutorials and books, a glossary of the field's terms, and a timeline of its history.

10 areas, 50 topics — and growing

Every topic has its own article — with the equations that matter and the papers they come from — linked to related topics, papers, groups and open problems.

5