🔢 mathematics

Hierarchical local null controllability for a non-degenerate quasi-linear parabolic system

This paper establishes the hierarchical local null controllability of a non-degenerate quasi-linear parabolic system on a bounded interval using a Stackelberg-Nash strategy with one leader and two followers, by first proving a Carleman inequality for the linearized adjoint problem and then applying Liusternik's Inverse Function Theorem.

Original authors: Cristian Loli, George J. Bautista, Juan Limaco, Rafael Lobosco, Dany Nina-Huaman, Luis P. Yapu

Published 2026-08-26

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Cristian Loli, George J. Bautista, Juan Limaco, Rafael Lobosco, Dany Nina-Huaman, Luis P. Yapu

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the world of engineering and physics, many natural processes unfold over time and space, spreading out like heat through a metal rod or a chemical spreading through water. Scientists describe these spreading processes using mathematical models called parabolic equations. A central challenge in this field is control: if you can influence a system in a specific region, can you steer it to a desired state, such as bringing all motion or concentration to a complete stop, by a specific deadline? This is known as controllability. Often, these systems are not simple; they involve multiple agents or decision-makers, each with their own goals. Sometimes these goals align, but often they compete. To manage such complex interactions, researchers use a strategy borrowed from economics and game theory, where one leader sets a broad plan and several followers react to it, each trying to optimize their own outcome. The followers reach a stable point where no one can improve their situation by changing their strategy alone, a state known as a Nash equilibrium. The leader then uses this stability to achieve a global objective.

The researchers in this study tackled a particularly difficult version of this problem. They looked at a system of two coupled equations that describe how two different quantities evolve and influence each other. What made this system hard to control was that the way the quantities spread depended on the quantities themselves. In simpler terms, the rules of the game changed as the game was being played. This "quasi-linear" nature creates significant mathematical obstacles because standard tools for controlling systems often fail when the system's behavior is tied to its own current state. The team focused on a scenario with three controllers: one leader acting on a main region, and two followers acting on separate, smaller regions. The leader's goal was to find a strategy that, once the two followers reacted optimally to reach their own stable balance, would drive both quantities in the system to exactly zero everywhere in the domain by a fixed final time.

The authors proved that this is possible, but only under specific conditions. They demonstrated that if the initial state of the system is close enough to zero—meaning the system starts with only a small disturbance—then a leader control exists that can successfully guide the entire system to a standstill. To reach this conclusion, the team first analyzed a simplified, linear version of the problem where the rules do not change. For this linear case, they developed a powerful mathematical tool, a type of weighted inequality, to show that the system could indeed be controlled to zero. This tool allowed them to estimate how the system's energy behaves and to prove that the leader's influence, combined with the followers' reactions, was sufficient to extinguish the system's activity.

Once the linear case was solved, the researchers faced the challenge of returning to the original, complex problem where the rules change with the state. They employed a sophisticated mathematical argument, essentially showing that because the system behaves predictably when it is very close to zero, the solution found for the simple linear case could be extended to the real, non-linear system for small disturbances. They established that the relationship between the controls and the system's outcome is smooth enough that a solution exists for the non-linear problem, provided the starting conditions are sufficiently small. The result is a rigorous proof that hierarchical control works for this class of difficult, self-interacting systems, but only when the system is not too far from a calm state. The study does not claim to solve the problem for large, chaotic initial states, but it firmly establishes that the hierarchical strategy is a viable and mathematically sound method for stabilizing these complex systems when they are near equilibrium.

Technical Summary: Hierarchical Local Null Controllability for a Non-Degenerate Quasi-Linear Parabolic System

Problem Statement
The manuscript addresses the hierarchical null controllability problem for a coupled system of non-degenerate quasi-linear parabolic equations posed on a bounded interval $I = (0, L)$ . The system involves two state variables, $y_1$ and $y_2$ , governed by equations where the diffusion coefficients $a(y_i)$ depend on the state variables themselves. The control mechanism employs a Stackelberg–Nash strategy involving three distinct controls:

A leader control ( $h$ ) acting on a subdomain $\omega$ .
Two follower controls ( $v_1, v_2$ ) acting on disjoint subdomains $\omega_1$ and $\omega_2$ .

The objective is to find a leader control $h$ such that, when the followers react optimally to minimize their respective cost functionals (reaching a Nash equilibrium), the resulting state of the system $(y_1, y_2)$ is driven to zero at a prescribed final time $T$ . The cost functionals for the followers measure the deviation of the state from desired trajectories in specific observation domains and penalize the magnitude of the control effort.

Methodology
The authors employ a multi-stage analytical approach to tackle the nonlinearity in the principal part of the differential operators:

Characterization of Nash Equilibrium: The authors first characterize the Nash quasi-equilibrium for the followers. By analyzing the convexity of the cost functionals (specifically for sufficiently large penalty parameters $\mu_i$ ), they establish that a Nash quasi-equilibrium corresponds to a Nash equilibrium. This leads to an optimality system coupling the state equations with adjoint equations.
Linearization: The nonlinear optimality system is linearized around the zero state. This step is crucial as the non-degenerate quasi-linear nature of the system creates significant analytical difficulties when nonlinearities appear in the principal part.
Carleman Estimates and Observability: For the linearized system, the authors derive a global Carleman inequality for the associated adjoint problem. This involves constructing specific weight functions that blow up at the initial and final times. These estimates are used to prove an observability inequality, which is the cornerstone for establishing the null controllability of the linearized hierarchical system.
Inverse Function Theorem: To extend the controllability result from the linearized system back to the original quasi-linear system, the authors utilize Lyusternik's Inverse Function Theorem in Banach spaces. This requires defining appropriate function spaces ( $Y$ and $Z$ ) with weighted norms to handle the regularity and the blow-up of the weight functions, and proving that the mapping associated with the system is continuously differentiable with a surjective derivative at the origin.

Key Contributions and Results

Existence of Local Null Controllability: The primary result (Theorem 2.1) establishes that for initial data $(f, g)$ and desired trajectories that are sufficiently small in the $H^1_0(I)$ and weighted $L^2$ norms, there exists a leader control $h$ and an associated Nash equilibrium $(v_1, v_2)$ such that the solution to the quasi-linear system satisfies $y_1(\cdot, T) = y_2(\cdot, T) = 0$ .
Handling Quasi-Linearities: The paper successfully addresses the challenge of nonlinearities in the principal part of the operator ( $-(a(y)y_x)_x$ ), a scenario that is more complex than standard semilinear cases where nonlinearities appear only in lower-order terms.
Hierarchical Framework: The work integrates game-theoretic concepts (Stackelberg–Nash) into the controllability of coupled quasi-linear parabolic systems, providing a rigorous mathematical framework for multi-agent control in distributed parameter systems.
Technical Estimates: The derivation of the observability inequality for the linearized adjoint system with specific weight functions and the subsequent application of Lyusternik's theorem constitute the core technical achievements.

Significance and Claims
The paper positions itself as a contribution to the growing field of hierarchical control for partial differential equations (PDEs). While previous works have extensively covered linear and semilinear systems, the authors note that the study of hierarchical controllability for quasi-linear parabolic systems remains comparatively unexplored due to the analytical difficulties posed by state-dependent diffusion coefficients.

The authors claim that their work represents a significant step in transitioning from approximate hierarchical strategies to rigorous controllability frameworks for nonlinear systems. By combining Carleman estimates with inverse mapping theorems, they demonstrate that hierarchical null controllability can be achieved even when the diffusion coefficients depend on the state variables, provided the initial data and target trajectories are sufficiently small. The results extend the applicability of Stackelberg–Nash strategies to general distributed parameter systems governed by quasi-linear parabolic equations, filling a gap in the literature where such nonlinearities appear in the principal part of the operator.

Limitations and Scope
The authors explicitly state that the controllability result is local. The existence of the control and the convergence to the zero state are guaranteed only for initial data and desired trajectories that are "sufficiently small" in the specified norms. This limitation is inherent to the use of Lyusternik's Inverse Function Theorem, which provides local surjectivity. The paper does not propose global controllability results or experimental validations, focusing strictly on the theoretical analysis of the mathematical model.

Technical Summary: Hierarchical Local Null Controllability for a Non-Degenerate Quasi-Linear Parabolic System

More like this