Ensemble Control of Dynamical Systems

Ensemble control addresses the challenge of steering large populations of dynamical systems, each characterised by varying parameters or initial conditions, with a single or limited set of control inputs. Such problems arise across diverse fields, from quantum spin ensembles and neural populations to arrays of mechanical devices and epidemiological models. The central difficulty lies in the inherent heterogeneity of the units, which precludes traditional feedback approaches for individual elements. Instead, ensemble control theory has developed open-loop methods and robust design principles that guarantee the collective behaviour follows desired trajectories or attains target states. Key strategies include optimal control formulations, polynomial approximation methods and iterative schemes based on Pontryagin’s Maximum Principle, enabling approximate yet precise manipulation of infinite or large finite ensembles. Recent advances have extended theoretical results on reachability and stabilisation to classes of affine, linear and nonlinear systems, while computational algorithms have been devised to approximate controls with increasing accuracy. Practical applications range from precise frequency-selective magnetic resonance ensembles and fault-tolerant quantum operations to synchronised stimulation of neural populations and coordinated actuation in micro-robotic swarms.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Research from all publishers

Efforts in optimal control theory have produced rigorous frameworks for steering ensembles of affine-control systems. A novel convergence approach replaces infinite ensembles with sequences of finite sub-ensembles, whose solutions approximate the original optimal control problem, and yields algorithmic schemes combining subspace projection of gradient fields with iterative Pontryagin-based updates. In neuroscience, control strategies for underactuated neural ensembles driven by optogenetic stimulation demonstrate how common inputs can induce complex spike-train patterns across heterogeneous integrate-and-fire neurons. By exploiting parameter variability, partial synchronisation and subensemble targeting become feasible, promising advances in sensory neuroprosthetics. Complementing these, polynomial approximation techniques have been developed to construct parameter-independent open-loop inputs for uniformly ensemble reachable linear systems. These methods distinguish between continuous-time and discrete-time dynamics, providing explicit computational procedures for single-input ensembles and illustrating the extension of discrete schemes to certain continuous models. Together, these works showcase a blend of theoretical depth and computational tractability that underpins emerging applications in quantum control, large-scale neural modulation and distributed mechanical systems.

Ensemble Control of Dynamical Systems publication trend

The graph below shows the total number of articles in ensemble control of dynamical systems across all publications each year (not limited to Nature Index journals).

Technical terms

Ensemble controllability: The capability of a single or limited control input to steer all members of a heterogeneous collection of dynamical systems to desired states.

Underactuated control: A scenario in which the number of available control inputs is insufficient to directly actuate every degree of freedom of the system.

Optimal control: A mathematical framework for determining control policies that minimise or maximise a predefined cost functional subject to system dynamics.

Pontryagin’s Maximum Principle: A set of necessary conditions for finding optimal trajectories in control problems, involving the Hamiltonian formulation and adjoint variables.

References

1. Optimal control of ensembles of dynamical systems*. ESAIM Control Optimisation and Calculus of Variations (2023).
2. Control strategies for underactuated neural ensembles driven by optogenetic stimulation. Frontiers in Neural Circuits (2013).
3. Polynomial methods to construct inputs for uniformly ensemble reachable linear systems. Mathematics of Control, Signals, and Systems (2023).