Burgers Control System Overview

• Burgers control systems are controlled PDEs based on viscous Burgers equations, outlining key principles of nonlinear convection–diffusion and diverse actuation methods.
• They employ a range of control architectures—including localized interior, boundary, and low-dimensional feedback controls—to achieve approximate, exact, and null controllability.
• Analytical and numerical studies reveal robust stabilization techniques alongside quadratic obstructions and nonlocal effects that shape the system's controllability properties.

Searching arXiv for recent and foundational papers on Burgers control systems to ground the article in the literature. Searching arXiv for recent and foundational papers on Burgers control systems to ground the article in the literature. Burgers control system denotes a family of controlled partial differential equations built on Burgers-type dynamics, most commonly the viscous equation

$y_t-y_{xx}+y\,y_x=\text{control},$

posed on a bounded interval or on $\mathbb R$, and supplemented by internal, boundary, scalar, distributed, or feedback actuation. In control theory, Burgers equations serve as a canonical nonlinear convection–diffusion model in which parabolic smoothing, nonlinear transport, nonlocality, and boundary effects can all be studied in a mathematically explicit setting. The literature exhibits a striking diversity of phenomena: exact and approximate controllability can hold for some actuation mechanisms, fail for others, and coexist with strong stabilization results; these outcomes depend sensitively on whether the control is spatially localized, spatially uniform, boundary-based, filtered through a nonlocal operator, or implemented through finite-dimensional feedback (Marbach, 2015, Shirikyan, 2016, Shirikyan, 2017, Araruna et al., 2024, Nguyen et al., 10 Jul 2025).

1. Controlled Burgers equations and actuation architectures

The prototypical controlled viscous Burgers system on a bounded interval is

$$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$

with homogeneous Dirichlet boundary conditions, where $\eta$ is the control input (Shirikyan, 2016, Shirikyan, 2017). This additive forcing formulation underlies much of the classical controllability and stabilization theory for Burgers-type systems. On $[0,1]$, localized interior control is often imposed through a support condition such as

$$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$

so that actuation is available only on a strict subinterval (Shirikyan, 2016). A different and highly singular architecture uses a spatially uniform scalar source

$y_t-y_{xx}+y\,y_x=u(t),$

where the control depends only on time and acts everywhere in space as a constant forcing term (Marbach, 2015, Nguyen et al., 10 Jul 2025). This model is central in the obstruction results discussed below.

Boundary control leads to another major branch of the theory. For generalized Burgers equations on $[0,1]$, one may impose

$y(t,0)=v(t),\qquad y(t,1)=0,$

while retaining a scalar interior forcing $u(t)$ that is uniform in space (Robin, 2022). In stabilization problems, Neumann boundary feedback laws are also used. For the shifted viscous Burgers equation

$\mathbb R$0

the boundary derivatives can be controlled by nonlinear feedbacks at both endpoints (Singh et al., 29 Nov 2025). Related Neumann feedback designs appear for Burgers equations with memory, where the control law also depends on a history variable (Singh et al., 1 Feb 2026).

Several important variants replace the standard transport velocity by a filtered quantity. The Burgers-$\mathbb R$1 model uses

$\mathbb R$2

so that the transport term becomes $\mathbb R$3 rather than $\mathbb R$4 (Araújo et al., 2020, Araruna et al., 2024). This creates a nonlocal regularized Burgers dynamics in which the control problem must be handled uniformly as $\mathbb R$5. More general convection laws are also studied, for instance

$\mathbb R$6

with a left boundary control $\mathbb R$7 and right Dirichlet condition (Robin, 2022). Beyond pure Burgers models, Burgers-type control systems include Burgers–Huxley equations with reaction terms (Singh et al., 2023, Patel et al., 30 Jun 2026), Korteweg–de Vries–Burgers equations with memory-type boundary control (Chentouf et al., 2023), and fluid–particle couplings in which Burgers dynamics interacts with a point mass subject to Newton’s law (Ramaswamy et al., 2020).

The choice of control space is equally varied. In some exact or null-controllability settings the control is distributed and square-integrable in space–time (Araruna et al., 2024), while in Agrachev–Sarychev-type approximate controllability one restricts to a finite-dimensional subspace, such as

$\mathbb R$8

yet still obtains strong reachability properties through nonlinear mode interactions (Shirikyan, 2017). On the real line, approximate controllability has even been achieved by an explicit 11-dimensional trigonometric control space (Shirikyan, 2013).

2. Controllability regimes and positive results

Controllability for Burgers systems is formulated in several distinct senses. Approximate controllability asks to steer the terminal state arbitrarily close to a target. Null controllability requires steering to zero. Exact controllability demands hitting the target exactly. Depending on the setting, the target may be an arbitrary profile, a trajectory of the uncontrolled system, or a constant state (Shirikyan, 2016, Shirikyan, 2017, Araújo et al., 2020).

A particularly influential positive result concerns the one-dimensional viscous Burgers equation with localized interior control. For

$\mathbb R$9

on $$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$0 with $$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$1, there exist positive constants $$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$2 and $$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$3 such that for any initial states $$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$4, one can choose a piecewise continuous control satisfying

$$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$5

This provides global exponential stabilization to trajectories, and when combined with a local exact controllability theorem of Fursikov–Imanuvilov yields global exact controllability to trajectories in finite time (Shirikyan, 2016). The same work proves that the time needed to enter an $$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$6-neighborhood is $$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$7, establishing a quantitative bridge between stabilization and exact controllability.

Approximate controllability by low-dimensional forcing is another major positive theme. For the 1D viscous Burgers equation on $$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$8,

$$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$9

approximate controllability at any time $\eta$0 is achieved with controls taking values in

$\eta$1

The same two-dimensional control space also yields a stronger property: simultaneous approximate controllability of the full state and exact controllability of finite-dimensional functionals (Shirikyan, 2017). On $\eta$2, a related Agrachev–Sarychev construction with

$\eta$3

where $\eta$4 are incommensurable, gives approximate controllability by an 11-dimensional trigonometric space in local topologies, without any decay assumption at infinity on the initial data (Shirikyan, 2013).

Null controllability can also be recovered in nonlocal regularized models. For the Burgers-$\eta$5 system

$\eta$6

with homogeneous Dirichlet conditions for both $\eta$7 and $\eta$8, local null controllability holds at any $\eta$9: there exists $[0,1]$0, independent of $[0,1]$1, such that every $[0,1]$2 with $[0,1]$3 can be driven to $[0,1]$4 at time $[0,1]$5 by a control $[0,1]$6, uniformly bounded independently of $[0,1]$7 (Araruna et al., 2024). The same work proves a large-time null-controllability criterion under $[0,1]$8, as well as convergence of controls and states to those of the classical Burgers equation as $[0,1]$9.

For the inviscid and viscous Burgers-$$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$0 systems with three scalar controls—one distributed control $$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$1 and two boundary controls $$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$2—the controllability picture is even stronger. The inviscid system is globally exactly controllable in $$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$3 for every $$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$4 and $$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$5, with control and state norms bounded uniformly in $$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$6 (Araújo et al., 2020). The viscous system is globally exactly controllable in $$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$7 to constant trajectories, again uniformly in $$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$8, by combining smoothing, approximate controllability, and local exact controllability around constant states (Araújo et al., 2020).

Generalized Burgers equations with both a scalar interior forcing and a boundary control admit small-time global null controllability under suitable flux exponents. For

$$\operatorname{supp}\eta \subset \mathbb R_+\times[a,b],$$9

the system is small-time global null controllable when $y_t-y_{xx}+y\,y_x=u(t),$0: for every $y_t-y_{xx}+y\,y_x=u(t),$1 and every $y_t-y_{xx}+y\,y_x=u(t),$2, there exist

$y_t-y_{xx}+y\,y_x=u(t),$3

such that $y_t-y_{xx}+y\,y_x=u(t),$4 (Robin, 2022). The same conclusion holds for the sign-flux version when $y_t-y_{xx}+y\,y_x=u(t),$5.

These positive results demonstrate that Burgers control systems can display robust exact, approximate, and null controllability, but only under control architectures compatible with the nonlinear transport structure. The literature also makes clear that such compatibility is not automatic.

3. Obstructions and non-controllability phenomena

Burgers control theory is equally notable for sharp negative results. The most prominent obstruction concerns the viscous Burgers equation with a spatially uniform scalar forcing,

$y_t-y_{xx}+y\,y_x=u(t),$6

posed on $y_t-y_{xx}+y\,y_x=u(t),$7 with homogeneous Dirichlet boundary conditions (Marbach, 2015, Nguyen et al., 10 Jul 2025). Although the equation is parabolic and therefore has infinite propagation speed, the system is not small-time locally null controllable around the zero equilibrium (Marbach, 2015). More precisely, there exist $y_t-y_{xx}+y\,y_x=u(t),$8 such that for every $y_t-y_{xx}+y\,y_x=u(t),$9 one can find $[0,1]$0 with $[0,1]$1 for which no control satisfying $[0,1]$2 drives the solution to zero at time $[0,1]$3 (Marbach, 2015).

The mechanism is genuinely nonlinear and second-order. The classical finite-dimensional Lie-bracket obstruction built on

$[0,1]$4

fails to detect it, because the corresponding second-order bracket vanishes formally at the origin (Marbach, 2015). The obstruction instead emerges through a quadratic expansion and a nonlocal coercive kernel acting on the control primitive. The decisive norm is not a classical $[0,1]$5-type quantity but the weaker $[0,1]$6 norm of the scalar control (Marbach, 2015). This indicates that even when parabolic smoothing is present and classical bracket tests vanish, a hidden one-sided drift can still prevent null controllability.

A later result strengthens this obstruction dramatically. For the same scalar-controlled system, local null controllability fails not only in small time but at every finite horizon: for every $[0,1]$7 there exists $[0,1]$8 such that for every $[0,1]$9 one can find arbitrarily small initial data $y(t,0)=v(t),\qquad y(t,1)=0,$0 with $y(t,0)=v(t),\qquad y(t,1)=0,$1 such that no control $y(t,0)=v(t),\qquad y(t,1)=0,$2 with

$y(t,0)=v(t),\qquad y(t,1)=0,$3

can steer the solution to zero at time $y(t,0)=v(t),\qquad y(t,1)=0,$4 (Nguyen et al., 10 Jul 2025). The proof uses a perturbative expansion, a carefully chosen even Fourier mode $y(t,0)=v(t),\qquad y(t,1)=0,$5, and a spectral multiplier $y(t,0)=v(t),\qquad y(t,1)=0,$6 whose strict negativity yields a coercive quadratic obstruction in the $y(t,0)=v(t),\qquad y(t,1)=0,$7 norm (Nguyen et al., 10 Jul 2025). This improves the earlier small-time obstruction of Marbach by ruling out finite-time local null controllability altogether.

Negative reachability results also occur under localized control. For the 1D viscous Burgers equation with control supported in $y(t,0)=v(t),\qquad y(t,1)=0,$8, global approximate controllability to arbitrary targets fails even if infinite control time is allowed: for any $y(t,0)=v(t),\qquad y(t,1)=0,$9 and any $u(t)$0, there exists $u(t)$1 such that every controlled trajectory remains at distance at least $u(t)$2 from $u(t)$3 for all $u(t)$4 (Shirikyan, 2016). The obstruction comes from an a priori bound on the solution in a portion of the domain outside the control region, showing that the reachable set is not dense in $u(t)$5.

An analogous failure of fixed-time approximate controllability appears for the generalized Burgers–Huxley equation with localized interior control. For every $u(t)$6, the interior-controlled system on $u(t)$7 is not approximately controllable in $u(t)$8, and the same conclusion holds for the boundary-controlled version (Patel et al., 30 Jun 2026). The proof employs a weighted energy estimate with

$u(t)$9

to show that targets with sufficiently large mass on a subinterval cannot be approximated in fixed time (Patel et al., 30 Jun 2026).

These negative results exclude a common misconception: parabolicity, diffusion, or nonlinear transport do not by themselves guarantee favorable local or global controllability. The geometry of the control operator, especially whether it excites enough spatial directions, is decisive.

4. Analytical mechanisms: return method, quadratic obstructions, and saturation

Three analytical paradigms dominate the modern theory of Burgers control systems: the return method, perturbative obstruction analysis, and the Agrachev–Sarychev saturation mechanism.

The return method is used when the linearization around the zero trajectory is not controllable. In the Burgers-$\mathbb R$00 inviscid system, local null controllability near the origin is obtained by constructing a nontrivial trajectory that returns from $\mathbb R$01 to $\mathbb R$02, linearizing around that path, proving controllability of the linearized equation, and closing the argument by a fixed-point theorem (Araújo et al., 2020). A similar high-level structure appears in the generalized Burgers small-time global null-controllability result, where the system is first driven toward a large nonzero steady state $\mathbb R$03, then steered back toward zero with a residual boundary layer, and finally brought exactly to zero using local parabolic controllability (Robin, 2022). In that setting, the return trajectory is encoded by the steady problem

$\mathbb R$04

whose boundary-layer structure is crucial to the proof (Robin, 2022).

Quadratic obstruction analysis is the central tool for proving non-controllability with scalar time-dependent forcing. In Marbach’s approach, one rescales time by $\mathbb R$05 and writes

$\mathbb R$06

where $\mathbb R$07 solves the controlled heat equation and $\mathbb R$08 solves a second-order equation driven by $\mathbb R$09 (Marbach, 2015). Projecting $\mathbb R$10 against a specific polynomial profile

$\mathbb R$11

one obtains a quadratic form

$\mathbb R$12

with asymptotic kernel

$\mathbb R$13

whose weakly singular structure leads to coercivity in $\mathbb R$14 on the primitive of the control, hence $\mathbb R$15 on the control itself (Marbach, 2015). In the later finite-time obstruction, the quadratic term is analyzed instead through time Fourier transform and explicit multipliers $\mathbb R$16 and $\mathbb R$17, in a strategy inspired by Coron, Koenig, and Nguyen’s work on KdV (Nguyen et al., 10 Jul 2025).

The saturation method, by contrast, explains why extremely low-dimensional controls can still yield approximate controllability. In the Burgers equation on $\mathbb R$18, the nonlinear transport operator mixes Fourier modes, and the iterative construction

$\mathbb R$19

satisfies

$\mathbb R$20

so repeated convexification expands the effective control space from $\mathbb R$21 to a dense union in $\mathbb R$22 (Shirikyan, 2017). On $\mathbb R$23, the corresponding mode-generation mechanism uses incommensurable frequencies $\mathbb R$24 and nonlinear trigonometric interactions to generate a dense countable frequency set (Shirikyan, 2013).

A plausible implication is that Burgers equations sit at a methodological crossroads within nonlinear PDE control: the same nonlinearity $\mathbb R$25 that generates dense mode cascades under distributed forcing can produce hard second-order obstructions when the control operator is too degenerate.

5. Feedback stabilization and closed-loop design

Beyond open-loop controllability, Burgers control systems admit a rich feedback stabilization theory. One line of work addresses finite-parameter feedback stabilization for Burgers-type models. For the “original Burgers’ equations” and a Burgers equation with nonlocal cubic nonlinearity, finite-dimensional feedback laws based either on low Fourier modes or on finitely many volume elements yield global exponential stabilization toward a concrete trajectory of the uncontrolled system (Gumus et al., 2019). In the original Burgers turbulence model, the controlled PDE–ODE system includes the term

$\mathbb R$26

where $\mathbb R$27 are the first Dirichlet eigenfunctions and $\mathbb R$28 is the feedback gain (Gumus et al., 2019). Under

$\mathbb R$29

the tracking error decays exponentially in $\mathbb R$30 (Gumus et al., 2019). For the nonlocal cubic model, sufficiently large $\mathbb R$31 and $\mathbb R$32 produce an arbitrary prescribed exponential convergence rate in $\mathbb R$33, and exponential stabilization also holds in $\mathbb R$34 (Gumus et al., 2019).

A second line uses Riccati-based feedback around steady states. For the controlled viscous Burgers system on $\mathbb R$35,

$\mathbb R$36

the target is a non-constant steady state $\mathbb R$37 (Akram, 2024). Writing $\mathbb R$38, the linearized operator is

$\mathbb R$39

with control operator $\mathbb R$40 (Akram, 2024). Under the coercivity condition

$\mathbb R$41

the semigroup generated by $\mathbb R$42 is analytic and exponentially stable in open loop up to finitely many unstable modes (Akram, 2024). After an exponential shift by a prescribed decay rate $\mathbb R$43, the algebraic Riccati equation

$\mathbb R$44

produces the optimal feedback

$\mathbb R$45

and the resulting closed-loop semigroup is exponentially stable (Akram, 2024). The same feedback stabilizes the original nonlinear perturbation locally by a Banach fixed-point argument (Akram, 2024).

Nonlinear Neumann boundary feedback yields global stabilization in still another architecture. For the shifted viscous Burgers equation

$\mathbb R$46

the controllers

$\mathbb R$47

are designed so that the boundary energy becomes dissipative (Singh et al., 29 Nov 2025). The discrete and continuous analyses both exploit the induced positive boundary terms. A related Lyapunov construction extends to Burgers equations with memory, where the controllers also depend on

$\mathbb R$48

and yield global stabilization in $\mathbb R$49, $\mathbb R$50, and $\mathbb R$51, as well as adaptive stabilization when $\mathbb R$52 is unknown (Singh et al., 1 Feb 2026).

These feedback results show that Burgers dynamics is amenable to both finite-dimensional and infinite-dimensional closed-loop synthesis. The preferred method depends on the actuation mechanism: modal damping is natural for internal low-mode feedback, Riccati design for localized distributed control around steady states, and Lyapunov boundary design for Neumann feedback.

6. Numerical control, discretization, and approximation of Burgers systems

The numerical analysis of Burgers control systems is itself a substantial subfield, because nonlinear convection, boundary feedback, and nonlocal terms complicate the preservation of stability and optimality under discretization.

For steady distributed optimal control governed by the one-dimensional Burgers equation,

$\mathbb R$53

with pointwise box constraints

$\mathbb R$54

piecewise linear finite elements for both state and control yield

$\mathbb R$55

under a second-order sufficient optimality condition (Rosero, 2014). If the optimal control is piecewise $\mathbb R$56, this improves to

$\mathbb R$57

which is the central convergence-rate result of that work (Rosero, 2014). The state and adjoint errors have the standard orders $\mathbb R$58 in $\mathbb R$59 and $\mathbb R$60 in $\mathbb R$61 (Rosero, 2014).

For optimal control of viscous Burgers on $\mathbb R$62,

$\mathbb R$63

particle methods provide a different discretization route. The control $\mathbb R$64 lies in

$\mathbb R$65

and the state is approximated by a distributional particle representation with mollified Dirac masses (Marburger et al., 2013). Under regularity assumptions, both the forward state and adjoint converge at order $\mathbb R$66, and a subsequence of discrete optimal controls converges strongly in $\mathbb R$67 to a continuous optimal control (Marburger et al., 2013).

In feedback stabilization, preserving exponential decay at the discrete level is crucial. For the nonlinear Neumann boundary-feedback problem discussed above, a $\mathbb R$68-scheme on a uniform mesh is proposed. It is conditionally stable for $\mathbb R$69 and unconditionally stable for $\mathbb R$70 (Singh et al., 29 Nov 2025). For $\mathbb R$71, the fully discrete solution satisfies exponential decay and first-order convergence of the state in discrete $\mathbb R$72, $\mathbb R$73, and $\mathbb R$74 norms, while the boundary controls also converge at first order (Singh et al., 29 Nov 2025). The discrete analysis relies on a special conservative discretization of the nonlinear term,

$\mathbb R$75

chosen so that a discrete boundary flux identity mirrors the continuous energy method (Singh et al., 29 Nov 2025).

The Riccati-stabilized localized-control problem around a non-constant steady state likewise admits a finite element approximation. The semidiscrete operator $\mathbb R$76 generates uniformly analytic semigroups, the discrete algebraic Riccati equation yields a uniformly exponentially stabilizing discrete feedback, and the stabilized state and control satisfy essentially quadratic-order error estimates: $\mathbb R$77

$\mathbb R$78

for any $\mathbb R$79 (Akram, 2024). Numerical experiments in that work confirm the theoretical decay and convergence rates (Akram, 2024).

For Burgers equations with memory under Neumann boundary feedback, $\mathbb R$80-conforming finite elements and the Ritz–Volterra projection yield optimal error estimates in $\mathbb R$81, $\mathbb R$82, and $\mathbb R$83 for the state, together with second-order convergence of the feedback controls (Singh et al., 1 Feb 2026).

This body of numerical work suggests that Burgers control systems form a favorable testbed for structure-preserving discretization: one can analyze not only state approximation, but also how stabilization mechanisms, feedback laws, and optimality systems survive numerical approximation.

7. Extensions, variants, and broader significance

The term “Burgers control system” encompasses more than the classical viscous equation. It includes nonlocal, generalized, stochastic, and coupled extensions that retain Burgers-type convection as a dominant structural feature.

The Burgers-$\mathbb R$84 model introduces a filtered velocity

$\mathbb R$85

which regularizes transport and enables uniform exact controllability results independent of $\mathbb R$86 (Araújo et al., 2020), while also admitting local null controllability with controls converging to those of the classical Burgers equation as $\mathbb R$87 (Araruna et al., 2024). This makes Burgers-$\mathbb R$88 a natural bridge between regularized and unregularized nonlinear transport control.

Fluid–structure interaction leads to Burgers–particle systems. In the one-dimensional model

$\mathbb R$89

with coupling $\mathbb R$90, a control acting only on the particle can steer the fluid velocity and particle velocity exactly to zero while bringing the particle position arbitrarily close to a prescribed target $\mathbb R$91, with a control time independent of the initial data (Ramaswamy et al., 2020). This is a global-in-data controllability result for a Burgers-type fluid–structure system.

Stochastic optimal control introduces noise and dynamic programming. For the stochastic Burgers equation on $\mathbb R$92,

$\mathbb R$93

with $\mathbb R$94 the Dirichlet Laplacian and $\mathbb R$95, the associated Hamilton–Jacobi–Bellman equation is an infinite-dimensional second-order integro-differential equation (Mohan et al., 2022). Semigroup smoothing, Bismut–Elworthy–Li formulas, and compactness arguments yield a mild HJB solution and a saturated negative-gradient feedback law $\mathbb R$96 (Mohan et al., 2022). This extends Burgers control theory into the Lévy-driven stochastic setting.

Burgers–Huxley control systems add reaction terms and can exhibit a different controllability geometry. Localized boundary feedback laws can stabilize the generalized Burgers–Huxley equation under Neumann actuation (Singh et al., 2023), while fixed-time approximate controllability fails for localized interior control, yet quasi-static approximate steering between steady states remains possible when the initial and target steady states lie in the same connected component of the steady-state set (Patel et al., 30 Jun 2026). This suggests that, in convection–reaction–diffusion systems, the steady-state manifold can replace exact reachability as the relevant control geometry.

Finally, boundary control of the Burgers equation linearized at a stationary shock reveals a singular perturbation aspect of Burgers control. For the linearized viscous problem on $\mathbb R$97 with control at the left endpoint, the null-controllability cost remains uniformly bounded as viscosity $\mathbb R$98 only above a threshold time $\mathbb R$99 satisfying

$$\partial_t u - \nu \partial_x^2 u + u\,\partial_x u = h(t,x) + \eta(t,x),$$00

The proof combines spectral analysis, moment methods, and complex analysis, and identifies an exponentially small eigenvalue associated with the shock profile as the source of the uniform-time threshold (Laheurte, 2024).

Taken together, these developments show that Burgers control systems occupy a central position in nonlinear PDE control. They are simple enough to permit explicit spectral, kernel, and energy analyses, yet rich enough to display most of the major phenomena of the field: low-dimensional approximate controllability, global exact controllability to trajectories, small-time global null controllability in suitable settings, sharp quadratic obstructions, finite-parameter tracking stabilization, Riccati-based feedback, adaptive boundary stabilization, stochastic dynamic programming, and delicate singular-limit behavior (Marbach, 2015, Shirikyan, 2016, Shirikyan, 2017, Araújo et al., 2020, Robin, 2022, Akram, 2024, Nguyen et al., 10 Jul 2025).