PaperlayerView exact revision on arXiv

Source-linked AI summary

Operator-Theoretic Stability and Observer Synthesis for Parameter-Dependent Vlasov--Maxwell Dynamics

Amadou Cissé, Mohamed Boutayeb

arXiv:2608.28349v1eess.SYmath.OC

TL;DR

The paper addresses the challenge of designing stable controllers and convergent observers for parameter-varying, infinite-dimensional Vlasov–Maxwell dynamics. It develops an operator-theoretic evolution framework with Lyapunov and H∞ inequalities, proves Galerkin consistency, and reports numerical convergence on a reduced benchmark.

  • Problem

    Stabilization, state estimation, and feedback design remain challenging for the infinite-dimensional, strongly coupled Vlasov–Maxwell system, and no parameter-dependent operator formulation had been proposed for it.

  • Method

    The paper models linearized dynamics as a parameter-dependent non-autonomous evolution and uses Lyapunov operators, operator differential LMIs, observer duality, H∞ conditions, and Galerkin projections.

  • Results

    The formulation establishes well-posedness, uniform growth properties, exponential stability, observer convergence, disturbance attenuation conditions, and Galerkin-consistent finite-dimensional LMIs, with reduced-benchmark simulations confirming predicted convergence.

  • Takeaways & Limitations

    The resulting operator and Galerkin formulations support numerical controller and observer synthesis while preserving the analytical structure of the Vlasov–Maxwell model.

  • Takeaways & Limitations

    The analysis is restricted to periodic spatial boundaries and does not address reflecting or absorbing boundary conditions.

Abstract

from arXiv · show

An operator--theoretic formulation is developed for the synthesis of parameter-dependent controllers and observers for the Vlasov--Maxwell system. The linearized dynamics are modeled as a non-autonomous evolution system whose generators depend on measurable plasma quantities. Well-posedness of the associated evolution family is established together with uniform growth bounds. Parameter-dependent Lyapunov operators yield operator differential LMIs ensuring uniform exponential stability and observer convergence. An $H_\infty$ extension provides disturbance attenuation conditions consistent with the intrinsic energy structure of the coupled Vlasov--Maxwell equations. Galerkin projections lead to finite-dimensional LMIs consistent with the operator inequalities, enabling reliable numerical synthesis while preserving the analytical structure of the original model. Numerical results on a reduced Vlasov--Maxwell benchmark confirm the predicted convergence properties.

1 INTRODUCTION

The paper extends parameter-dependent operator synthesis to the coupled Vlasov–Maxwell system, addressing stability and state estimation for its infinite-dimensional, strongly coupled dynamics. It establishes well-posedness, Lyapunov and H∞ conditions, Galerkin consistency, and numerical validation.

  • Stabilization, state estimation, and feedback design remain challenging because Vlasov–Maxwell dynamics are infinite-dimensional and strongly couple kinetic and electromagnetic components.
  • No parameter-dependent operator formulation had previously been proposed for Vlasov–Maxwell equations, despite time-varying macroscopic quantities influencing the linearized dynamics.
  • Macroscopic quantities such as kinetic energy, field energy, and current-density moments provide physically motivated scheduling variables through their effects on plasma frequency, temperature, and pressure terms.
  • The framework extends existing parameter-dependent synthesis to combine phase-space transport, current-induced electromagnetic coupling, and parameter-varying field dynamics.
  • Well-posedness, uniform growth bounds, exponential stability, observer convergence, H∞ robustness, and Galerkin consistency are established, with reduced-model simulations confirming predicted convergence.

2 MODEL AND PARAMETER-DEPENDENT FORMULATION

The paper formulates controlled relativistic Vlasov–Maxwell dynamics on a periodic spatial domain with kinetic, electric, and magnetic state variables. It uses weighted kinetic spaces, energy-based electromagnetic operators, perturbation linearization, and measurable parameter scheduling to obtain an operator representation with observable outputs.

  • 2 MODEL AND PARAMETER-DEPENDENT FORMULATION: The model evolves the particle distribution f(t, x, v) together with electric and magnetic fields on a periodic spatial domain and unbounded velocity space.
  • 2.1 Controlled Vlasov–Maxwell equations: External actuation enters through the electric-field equation, preserving Maxwell coupling while representing sources such as coils, antennas, or distributed current actuation.
  • 2.1 Controlled Vlasov–Maxwell equations: The analysis is restricted to periodic boundaries and does not address reflecting or absorbing boundaries, where boundary-induced singularity or concentration effects may arise.
  • 2.2 Functional Setting: The kinetic space is weighted because the unbounded velocity domain makes velocity moments unsuitable for a plain unweighted space.
  • 2.2 Functional Setting: The Maxwell operator is constructed on H(curl; Ω) × H(curl; Ω), with boundary conditions eliminating curl integration-by-parts terms and yielding a skew-adjoint structure.
  • 2.2 Functional Setting: The electromagnetic state uses an energy inner product equivalent to the standard product L2 norm, with adjoints and skew-adjointness defined in that energy space.
  • 2.3 Nominal trajectory and perturbation variables: Perturbations are defined around a nominal controlled solution, and linearization couples phase-space transport with electromagnetic perturbations through Lorentz-force and Maxwell terms.
  • 2.4 Parameter scheduling and parameter-dependent operator structure: A measurable parameter trajectory schedules frozen operating regimes, while the resulting operators encode transport, Lorentz coupling, Maxwell dynamics, current functionals, and macroscopic observations.

3 WELL-POSEDNESS OF THE PARAMETER-DEPENDENT VLASOV– MAXWELL SYSTEM

The paper establishes well-posedness for the parameter-dependent Vlasov–Maxwell evolution by combining common-domain operator assumptions, Kato stability, semigroup generation, and bounded coupling perturbations. The resulting evolution family has a unique mild solution and bounds uniform across admissible parameters.

  • The operator family A(ρ) models linearized kinetic–electromagnetic dynamics driven by a measurable scheduling parameter, and the section establishes evolution existence, uniqueness, and regularity.
  • The assumptions require a parameter-independent dense domain, semigroup generation for each frozen operator, uniform Kato stability, continuous parameter dependence, piecewise C1 trajectories, and bounded observation-related families.
  • The nominal fields are time-independent for fixed ρ, so non-autonomous behavior arises exclusively from the varying parameter trajectory ρ(·).
  • The transport operator retains the full first-order kinetic structure on its natural graph domain because the Lorentz transport term is itself unbounded.
  • The Maxwell block is skew-adjoint and generates a unitary C0-group, while the kinetic and electromagnetic blocks together generate a strongly continuous semigroup.
  • The remaining linearized coupling is treated as a bounded perturbation under regularity and decay assumptions, yielding semigroup generation for each fixed parameter.
  • For every admissible parameter trajectory and L2_loc input, a unique mild solution exists through a strongly continuous evolution family.
  • The evolution-family growth constants are uniform in ρ, providing the analytical foundation for later Lyapunov and H∞ estimates.

4 PARAMETER-DEPENDENT LYAPUNOV FUNCTIONALS AND OPERATOR DIFFERENTIAL LMIS

The section develops parameter-dependent Lyapunov functionals and operator differential LMIs for uniform exponential stability, then establishes consistency of their Galerkin approximations.

  • Parameter-dependent Lyapunov functionals: Bounded, self-adjoint, coercive Lyapunov operators P(ρ) are assumed continuously differentiable with bounded parameter derivatives.Their parameter dependence permits the Lyapunov functional to track scheduling variations.
  • Parameter-dependent Lyapunov functionals: The derivative of the Lyapunov functional separates frozen-parameter dynamics from terms caused by scheduling-parameter variation.The chain rule introduces the contribution involving the derivative of P(ρ).
  • Operator differential LMIs: The operator differential LMI incorporates parameter variations through ν_i ∂ρ_iP(ρ) and extends the classical operator Lyapunov inequality.When P(ρ) is constant, the standard operator Lyapunov inequality is recovered.
  • Operator differential LMIs: If the operator differential LMI holds, every solution along admissible parameter trajectories is uniformly exponentially stable.The result follows by combining the Lyapunov derivative inequality with Grönwall’s inequality.
  • Galerkin discretization: Galerkin projections produce finite-dimensional LMIs whose feasibility and consistency preserve the link to the infinite-dimensional operator inequalities.Under uniform coercivity, bounded gains, and strong projection convergence, feasible projected LMIs yield limiting operators satisfying the operator inequality.

5 PARAMETER-DEPENDENT OBSERVER DESIGN

The section extends Lyapunov-based synthesis to parameter-dependent observers, providing dual operator LMIs for exponential estimation-error convergence and consistent finite-dimensional observer design.

  • Observer formulation: A parameter-dependent Luenberger observer uses a continuously parameterized gain L(ρ) to reconstruct the state from measured outputs.The estimation error is defined as e = X−ˆX and analyzed through its induced error dynamics.
  • Dual operator differential LMI: The dual operator differential LMI replaces the controller feedback term B(ρ)K(ρ) with the observer injection term L(ρ)C(ρ).This gives the observer condition the same Lyapunov structure as the primal controller inequality.
  • Observer convergence: If the dual operator differential LMI holds, the estimation error has a unique mild solution and converges uniformly exponentially along admissible parameter trajectories.The guarantee applies when the parameter derivative remains in the admissible set.
  • Galerkin observer synthesis: Projected dual LMIs require vertex verification when the relevant operators depend affinely on ρ, reducing observer synthesis to finite-dimensional convex constraints.The gain is recovered as L(ρ) = Q_o(ρ)^−1Y_o(ρ).
  • Galerkin observer synthesis: Uniformly coercive projected observer solutions with bounded recovered gains yield limiting operators satisfying the dual operator inequality and the infinite-dimensional convergence estimate.The consistency argument uses weak-⋆ compactness and passage to the limit in quadratic forms.

6 H∞PERFORMANCE AND ROBUST PARAMETER-DEPENDENT CONTROL

The section formulates H∞ disturbance attenuation through an operator differential inequality and shows that Galerkin LMIs preserve this condition, including for dual observer synthesis.

  • H∞ control formulation: The H∞ formulation models disturbances through B_w(ρ) and evaluates performance outputs such as electromagnetic energy, current deviations, or selected kinetic moments.The design seeks a prescribed disturbance attenuation level γ for zero initial conditions.
  • Operator H∞ inequality: The operator H∞ inequality expresses a differential energy balance for the closed-loop parameter-dependent dynamics.A coercive parameter-dependent Lyapunov operator characterizes dissipativity along trajectories.
  • H∞ guarantees: Satisfaction of the operator inequality ensures uniform exponential stability for zero disturbance and the disturbance attenuation property for zero initial condition.The stability conclusion uses coercivity, while attenuation follows by integrating the energy inequality.
  • Galerkin approximation: Galerkin LMIs provide consistent finite-dimensional approximations of the operator H∞ condition, with operator feasibility implying feasibility of all Galerkin LMIs.Conversely, uniformly coercive feasible sequences admit subsequential limit operators satisfying the operator condition in quadratic-form sense.
  • Dual H∞ observer formulation: The dual H∞ inequality yields observer gains that provide exponential error decay without disturbance and attenuation at level γ from disturbance to the selected error output.The primal and dual inequalities support coordinated controller–observer synthesis.

7 JOINT

The joint synthesis designs parameter-dependent feedback and observer gains through operator differential LMIs for uniform stability and H∞ performance. Galerkin projections preserve this operator-level structure and support consistent finite-dimensional synthesis.

  • Joint controller–observer synthesis: The formulation jointly designs parameter-dependent feedback gain K(ρ) and observer gain L(ρ) for the Vlasov–Maxwell system.It adapts a separation-type argument to the non-autonomous, infinite-dimensional setting.
  • Joint operator H∞ condition: A common operator inequality with coercive Lyapunov operators ensures augmented-system stability and an H∞ bound for disturbances.For w ≡ 0, coercivity yields uniform stability; with zero initial condition, integration gives the H∞ estimate.
  • Galerkin realization: Galerkin projection converts the joint operator inequality into finite-dimensional LMIs whose convex variables recover the feedback and observer gains.The projected operators define the finite-dimensional synthesis problem.
  • Galerkin consistency: If finite-dimensional joint LMIs remain feasible with uniformly coercive matrices and bounded gains, limiting coercive operators and bounded gains satisfy the joint operator inequality.The consistency argument uses strong convergence of projections and weak-⋆ compactness of extended Galerkin solutions.
  • Galerkin consistency: Uniform feasibility of primal, dual, and joint Galerkin LMIs implies that the limiting controller–observer pair satisfies the operator-level H∞ condition.This establishes consistency between numerically synthesized gains and the infinite-dimensional model.
  • Joint controller–observer synthesis: Joint controller–observer synthesis extends the separation principle to parameter-dependent Vlasov–Maxwell models through energy-consistent operator inequalities.The construction combines controller and observer design within the same operator-level framework.

8 NUMERICAL RESULTS

The reduced benchmark evaluates parameter-dependent L2 and H∞ observer synthesis under computationally tractable discretizations. Both designs produce stable simulations and convergence of electromagnetic and kinetic state estimates, while larger phase-space discretizations rapidly increase computational cost.

  • Computational scope: The numerical section is a proof-of-concept validation because semidefinite-program complexity makes large-scale phase-space discretizations rapidly intractable.The state dimension scales as n = NxNv + 2Nx, and standard interior-point methods typically have O(n6) complexity.
  • Benchmark setup: The benchmark uses a one-dimensional spatial setting, one velocity variable, periodic boundaries, and four Hermite velocity modes to preserve tractability.The truncation (Nx, Nv) = (6, 4) produces a reduced model of dimension n = 48.
  • Observer designs: The measured output is the electric field, and observer synthesis compares L2 decay-rate maximization with H∞ disturbance-to-error gain minimization.The SDP also imposes normalization, soft upper bounds, and injection-operator constraints to reduce ill-conditioning.
  • Numerical outcomes: β = 1.21 × 10−5 for L2, while H∞ yields β = 7.82 × 10−6 and γ = 1.57 × 10−8.Despite optimal_inaccurate solver status, both observers produce stable, well-conditioned simulations; the empirical H∞ attenuation is γemp ≈ 6.26.
  • State reconstruction: Figures 1–4 show convergence of electric-field estimates, kinetic errors, and phase-space reconstructions for the observer designs.The phase-space snapshots recover both spatial modulation and velocity structure, while the error plots track electromagnetic and kinetic norms over time.
  • Error analysis: ∥E −ˆE∥L2x ≈10−3 and ∥f −ˆf∥L2x,v ≈3 × 10−2, with L2 asymptotic rates λE = −2.39 × 10−2 and λf = −9.56 × 10−2.The reported rates confirm exponential convergence with distinct robustness–performance trade-offs.

9 CONCLUSION

The paper develops an operator-dependent control and observer framework for Vlasov–Maxwell dynamics and validates it numerically on a reduced benchmark. The framework combines stability, observer convergence, disturbance attenuation, and Galerkin-consistent finite-dimensional synthesis.

  • The paper models linearized Vlasov–Maxwell dynamics as a non-autonomous evolution system with well-posedness and uniform growth properties.
  • Parameter-dependent Lyapunov operators produce operator differential LMIs for exponential stability, observer convergence, and H∞ disturbance attenuation.
  • Galerkin projections yield finite-dimensional LMIs consistent with the underlying operator inequalities, supporting numerical synthesis on a reduced benchmark.

A AUXILIARY RESULTS USED IN THE PROOFS

The appendix identifies the functional-analytic results used to justify well-posedness and Galerkin consistency in the paper’s analysis.

  • The appendix recalls functional-analytic results explicitly used in the well-posedness and Galerkin-consistency arguments.

A.1 Kato’s theorem for non-autonomous evolution equations

Kato’s theorem provides existence, uniqueness, and growth bounds for the non-autonomous evolution family under uniform generator assumptions. The recalled results also include Stone’s unitary-group characterization and Banach–Alaoglu weak-⋆ compactness.

  • A common-domain family of closed densely defined operators generates a unique evolution family when each operator generates a C0-semigroup, resolvent bounds are uniform, and t 7→A(t)x is continuous.
  • The evolution family satisfies S(t, s)S(s, r) = S(t, r), S(s, s) = I, and ∥S(t, s)∥≤Meω(t−s), with unique mild solutions.
  • Stone’s theorem characterizes strongly continuous one-parameter unitary groups through unique self-adjoint generators and conversely constructs such groups from self-adjoint operators.
  • Banach–Alaoglu supplies weak-⋆ compactness of bounded subsets of a dual space, including the closed unit ball.
Loading 2608.28349v1…