Simultaneous Stochastic Resetting

• Simultaneous stochastic resetting is a protocol where the entire state of a system is returned to a preset configuration at random times.
• It produces nonequilibrium stationary states by inducing global correlations and controlling collective dynamics in diverse many-body systems.
• Its mathematical framework, based on renewal equations, unifies resetting in extended dynamical systems and optimizes processes like synchronization and search.

Simultaneous stochastic resetting is a reset protocol in which a single random event acts jointly on a high-dimensional state, sending all particles, fields, oscillators, or other degrees of freedom back to a prescribed configuration. In interacting-particle language this is global resetting, in contrast to local resetting, where different constituents reset independently (Nagar et al., 2023). In configuration-space form, it is the case in which an entire state $x_n$, $\boldsymbol X(t)$, or $\mathcal C(t)$ is restored at common reset times rather than coordinate by coordinate (Aron et al., 2024, Keidar et al., 2024, Evans et al., 2019). Recent work has established simultaneous resetting as a unifying mechanism for generating nonequilibrium stationary states, dynamically emergent correlations, synchronization, first-encounter optimization, and control of spatiotemporal chaos in many-body systems (Mauro et al., 8 Sep 2025, Sarkar et al., 2022, Suarez-Jimenez et al., 22 Apr 2026).

1. Definition and representative realizations

In the most direct formulation, simultaneous stochastic resetting means that at random times the whole state is reset at once. For a single map this is simply $x_n\to x_0$, whereas for a lattice it means $x_{n,i}\to x_{0,i}$ for all sites $i$ at the same time (Aron et al., 2024). In the general adaptive-resetting framework, the same idea is expressed by treating the joint state of all constituents as one high-dimensional variable $\boldsymbol X$, together with a single reset hazard $r(\boldsymbol X,t)$ that governs global reset events (Keidar et al., 2024). The interacting-particle resetting review makes the same distinction explicitly: simultaneous resetting corresponds to global resetting, while independent constituent-wise resets correspond to local resetting (Nagar et al., 2023).

System	Reset object	Principal consequence
Brownian gas	All particle positions reset to the origin	Dynamically emergent correlations and correlated NESS (Mauro et al., 8 Sep 2025)
Coupled chaotic maps	Whole lattice configuration reset to $x_0$	Simultaneous suppression of Lyapunov exponent and butterfly velocity (Aron et al., 2024)
Kuramoto oscillators	All phases reset to a fixed initial configuration	Reset-induced synchronized or pseudo-synchronized stationary state (Sarkar et al., 2022)
Multiple walkers on a network	All walkers reset to their respective initial nodes	Modified mean first-encounter times and synchronization–exploration trade-off (Suarez-Jimenez et al., 22 Apr 2026)
Interacting particle systems	Entire configuration $\mathcal C$ reset to $\boldsymbol X(t)$0	Global NESS, currents, and reset-controlled phase structure (Nagar et al., 2023)

The concept extends beyond explicitly many-body physical systems. The adaptive-resetting formulation emphasizes that any multidimensional process can be reset as a single object, so simultaneous resetting also covers high-dimensional molecular configurations and other collective state variables (Keidar et al., 2024). A broader implication is that “simultaneous” refers primarily to the reset operator acting on a joint state, not to any specific microscopic interaction.

2. Mathematical structure

For Poissonian global resetting of a full configuration $\boldsymbol X(t)$1 to $\boldsymbol X(t)$2 at rate $\boldsymbol X(t)$3, the configuration-space renewal equation is

$\boldsymbol X(t)$4

with stationary measure

$\boldsymbol X(t)$5

This is the natural many-body analogue of the standard single-particle renewal equation and makes explicit that simultaneous resetting is a renewal process in configuration space (Evans et al., 2019).

For discrete-time extended dynamical systems, the same structure appears as

$\boldsymbol X(t)$6

where $\boldsymbol X(t)$7 is the full lattice state. Here the reset event replaces the entire configuration by its initial value $\boldsymbol X(t)$8 (Aron et al., 2024). This is a prototypical implementation of simultaneous stochastic resetting in an extended classical system.

The adaptive-resetting formalism generalizes the same mechanism to state- and age-dependent protocols. It introduces a reset hazard

$\boldsymbol X(t)$9

so that the reset time depends on the joint trajectory of a high-dimensional process (Keidar et al., 2024). In that framework, simultaneous resetting of multiple constituents is realized by taking $\mathcal C(t)$0 to be the joint state and applying one global reset rule. The same paper derives a general mean-first-passage formula,

$\mathcal C(t)$1

which remains valid even when the reset time $\mathcal C(t)$2 is path-dependent (Keidar et al., 2024).

A plausible implication is that several apparently different simultaneous-reset models are mathematically unified by two ingredients: a common renewal clock and a reset map acting on the full state. The specific phenomenology then comes from the bare dynamics between resets.

3. Nonequilibrium stationary states and dynamically emergent correlations

A central many-body consequence of simultaneous resetting is the generation of correlations among otherwise non-interacting constituents. For $\mathcal C(t)$3 independent Brownian particles with common reset times, the stationary joint distribution is

$\mathcal C(t)$4

where $\mathcal C(t)$5 is the stationary distribution of the time since the last reset. This has a conditionally independent and identically distributed (CIID) structure: conditional on the common age $\mathcal C(t)$6, the particles are IID, but averaging over $\mathcal C(t)$7 produces strong collective correlations (Mauro et al., 8 Sep 2025). The same work shows that the existence of the stationary state requires a finite mean inter-reset time and that observables such as density, extreme values, gap statistics, and full counting statistics can then be computed exactly.

The effect persists in confinement, but confinement changes the correlation structure in a geometry-dependent way. For $\mathcal C(t)$8, with exact analysis for harmonic confinement and box confinement, the stationary state is controlled by the competition between the confinement length and the resetting length (Mauro et al., 30 Jun 2026). In both cases the density crosses over between confinement-dominated and resetting-dominated regimes, but the normalized correlation coefficient behaves differently: in box confinement it is non-monotonic and overshoots the unconfined value, whereas in harmonic confinement it increases monotonically toward the unconfined limit (Mauro et al., 30 Jun 2026). For general $\mathcal C(t)$9, the behavior is monotonic for $x_n\to x_0$0 and non-monotonic for $x_n\to x_0$1, with

$x_n\to x_0$2

The same paper identifies three EVS universality classes for $x_n\to x_0$3, $x_n\to x_0$4, and the singular limit $x_n\to x_0$5 (Mauro et al., 30 Jun 2026).

An important conceptual point is that one-point observables need not reveal the presence of global correlations. In the confined Brownian gas, the density may look similar in harmonic and hard-wall confinement, while correlation-sensitive observables and edge statistics differ sharply (Mauro et al., 30 Jun 2026). This directly refutes the common simplification that simultaneous resetting is equivalent to merely changing the single-particle marginal.

4. Control of collective dynamics and dynamical phase transitions

Simultaneous resetting has been developed as a control mechanism for collective nonlinear dynamics. In a coupled logistic-map lattice with global resetting of the entire state to its initial configuration, the Lyapunov exponent and butterfly velocity are both reduced by the reset probability $x_n\to x_0$6 (Aron et al., 2024). For the single-map problem,

$x_n\to x_0$7

while in the lattice the reset-renormalized butterfly velocity is

$x_n\to x_0$8

The paper identifies a dynamical phase transition at which the Lyapunov exponent and butterfly velocity vanish simultaneously, arresting spatiotemporal chaos and ballistic information spreading (Aron et al., 2024).

A closely related control effect appears in the Kuramoto model with global resetting. There, all oscillators are reset simultaneously at Poissonian times of rate $x_n\to x_0$9 to a fixed initial phase configuration with order parameter $x_{n,i}\to x_{0,i}$0 (Sarkar et al., 2022). In the Lorentzian-frequency case, the bare model has the standard synchronization threshold $x_{n,i}\to x_{0,i}$1, but resetting reshapes the phase diagram. For $x_{n,i}\to x_{0,i}$2, the critical line

$x_{n,i}\to x_{0,i}$3

separates a regime in which the stationary distribution $x_{n,i}\to x_{0,i}$4 is peaked near $x_{n,i}\to x_{0,i}$5 from a regime in which it develops a peak at finite $x_{n,i}\to x_{0,i}$6, a reset-induced pseudo-synchronized phase (Sarkar et al., 2022). For $x_{n,i}\to x_{0,i}$7, the threshold

$x_{n,i}\to x_{0,i}$8

controls whether the stationary distribution is concentrated near the bare synchronized value or suppressed there in favor of the reset-imposed coherence (Sarkar et al., 2022).

These two examples establish a common theme: simultaneous resetting can act as a coarse global intervention that competes with intrinsic many-body relaxation. Depending on the observable, the result is either a transition to localization and arrested spreading or a transition to reset-maintained order.

5. Search, encounters, and optimization

For multiple random walkers on networks, simultaneous resetting means that all walkers are synchronously returned to their respective initial nodes with probability $x_{n,i}\to x_{0,i}$9 at each time step (Suarez-Jimenez et al., 22 Apr 2026). The collective transition matrix is the tensor product of the individual transition matrices, while resetting adds a rank-one operator that sends the full configuration to the reset configuration. In this setting, the mean first-encounter time (MFET) to a node $i$0 is defined as the mean time for all walkers to meet for the first time at $i$1 (Suarez-Jimenez et al., 22 Apr 2026). The paper derives exact spectral formulas for the MFET in terms of the eigenvalues and eigenvectors of the no-reset transition operator.

A general criterion determines when a nonzero simultaneous reset probability is beneficial. If $i$2 denotes the coefficient of variation of the no-reset first-encounter time, then resetting lowers the MFET when

$i$3

and the optimal reset probability $i$4 satisfies

$i$5

The same work shows that simultaneous resetting is more efficient than independent resetting on homogeneous networks, while independent resetting can outperform it on heterogeneous networks, revealing a trade-off between synchronization and exploration (Suarez-Jimenez et al., 22 Apr 2026).

The adaptive-resetting framework broadens this optimization viewpoint. Because the joint state $i$6 can represent multiple walkers, collective variables, or an entire molecular configuration, simultaneous resetting can be analyzed through reweighting from reset-free trajectories and optimized by machine learning for observables such as first-passage times or stationary distributions (Keidar et al., 2024). The paper emphasizes that this eliminates exhaustive brute-force sampling across protocols and directly accommodates high-dimensional global resets, including molecular-dynamics configurations (Keidar et al., 2024).

A plausible implication is that simultaneous resetting is especially natural for coordinated search and restart problems: the reset is no longer an individual restart, but a synchronization event acting on the full collective trajectory.

6. Distinctions, neighboring concepts, and broader implications

The first conceptual distinction is between simultaneous/global and independent/local resetting. In interacting-particle systems this difference is structural, not merely terminological: global resetting replaces the entire configuration $i$7, whereas local resetting acts on one particle or one degree of freedom at a time (Nagar et al., 2023). In network encounter problems, the distinction translates into synchronization versus exploration: shared resets correlate walkers, independent resets decorrelate them (Suarez-Jimenez et al., 22 Apr 2026).

A second distinction is between fully simultaneous resets and more general correlated resetting events. The network-growth model with node deletion can be framed as a resetting problem in which resetting one node to degree zero simultaneously decreases the degrees of all its neighbors (Artime, 2022). One reset event thereby alters $i$8 degrees at once. This is not a full reset of the whole network configuration, but it is a solvable example of many-body resetting with correlated transitions (Artime, 2022). It shows that simultaneous stochastic resetting belongs to a wider class of reset protocols in which the reset event is itself collective.

A third distinction concerns stationarity versus ergodicity. More broadly, stochastic resetting can render a process stationary while leaving it non-ergodic, as shown for geometric Brownian motion under resetting (Stojkoski et al., 2021). This suggests that simultaneous resetting should not be identified automatically with equilibration: global resets can stabilize a nonequilibrium stationary state, but they need not restore equilibrium behavior.

The general stochastic-jump representation of resetting as a jump-diffusion process provides a useful formal backdrop. For Brownian motion with Poissonian resetting to $i$9,

$\boldsymbol X$0

with $\boldsymbol X$1 a Poisson process of intensity $\boldsymbol X$2, the reset is represented as a jump term to the reset state (Magdziarz et al., 2023). A plausible extension is that simultaneous resetting of a vector-valued state is captured by replacing the scalar jump by a common jump process acting on all coordinates at once. This interpretation is consistent with the configuration-space renewal equations used throughout many-body resetting.

Taken together, the current literature presents simultaneous stochastic resetting as a general many-body restart mechanism with three defining features: a common reset clock, a reset map acting on the joint state, and a competition between reset-induced synchronization and the intrinsic dynamics between resets. Its principal outcomes are reset-dependent nonequilibrium stationary states, dynamically emergent collective correlations, and sharp modifications of transport, synchronization, information spreading, and search efficiency across classical and, in broader review contexts, quantum systems (Evans et al., 2019, Nagar et al., 2023).