Quasi-Linear Conservation Laws

• Quasi-linear systems of conservation laws are mathematical models that describe wave propagation and continuum mechanics by encoding physical invariants like mass, momentum, and energy.
• They integrate algebraic and geometric structures, enabling classification through symmetry, duality, and nonlinear self-adjointness to reveal invariant conservation properties.
• Modern approaches leverage variational and numerical methods, symmetry constraints, and integrability techniques to analyze stability, shocks, and singular behavior in these systems.

Quasi-linear systems of conservation laws are foundational in the analysis of nonlinear partial differential equations modeling wave propagation, continuum mechanics, fluid dynamics, and related areas. These systems take the form

$\partial_t u + \partial_x f(u) = 0$

where $u(x, t)$ is a vector of conserved variables and $f(u)$ is the (typically nonlinear) flux. The “quasi-linear” designation emphasizes the dependence of the system coefficients on the unknown function itself, rather than strict linearity. Such systems are often hyperbolic, potentially admit discontinuous weak solutions (e.g., shocks), and possess deep connections to symmetry, geometry, variational principles, and integrability.

1. Algebraic and Geometric Structure

Quasi-linear systems frequently encode physical invariants (mass, momentum, energy) and admit profound geometric interpretations. For systems with constraints, as in multiphase flow, the configuration space of solutions may be a manifold $M$ defined implicitly by constraints $C(u) = 0$ within the ambient space $\mathbb{R}^m$ (Reintjes, 2015). The dynamics are then described intrinsically on $M$: $$g(w)_t + f(w)_x = 0,\quad w \in M \subset \mathbb{R}^m$$ with $g$ and $f$ the accumulation and flux functions restricted to the manifold.

The geometric approach extends standard solvability results: strictly hyperbolic systems defined on $M$ retain classical wave-interaction behavior. Eigenvalue problems and wave decomposition are reinterpreted via the tangent bundle of $M$: $df(w)(T) = \lambda\, dg(w)(T)$ where $T$ is a tangent vector to $M$, so wave families and their interaction remain meaningful in the constrained setting.

The projective geometry of conservation laws further enables classification and duality. For one-dimensional systems with additional conservation laws, the associated ruled hypersurface in projective space codifies the system’s invariants and characterizes duality relations, autodual Hamiltonian structures, and Temple-class systems (Agafonov, 2019).

2. Symmetry and Canonical Form

The presence of a nontrivial symmetry group imposes strong constraints on the admissible structure of quasi-linear systems. Imposing invariance under a symmetry group $S_R$ on the system’s entropy extension (potential) $\psi(z)$ requires that the lifted transformation preserve the form of the system: $\psi(z)^* = (X + c_Z I_m)\psi(z)$ for particular matrices $X, Z$ and constant $c_Z$ (Sever, 25 Sep 2025). This constraint often forces $\psi(z)$ to be of separable form,

$\psi(z) = \xi(z)\zeta_z(z)^T$

with implications for the structure of $a(z)$, the flux function.

Notably, when the symmetry group is sufficiently rich (maximal), the only systems supporting invariant entropy extensions with physically meaningful vanishing-dissipation limits in high dimensions are Euler-type systems (isentropic or extended forms incorporating conservation of energy, temperature, etc.). Thus, incorporating built-in symmetries directly into the conservation law system canonically restricts the class of models to physically robust frameworks in multi-dimensional fluid dynamics.

3. Construction and Classification of Conservation Laws

Modern methods extend Noether’s theorem to the non-variational setting via nonlinear self-adjointness. For a differential equation or system

$F(x, u, u_1, u_2, \ldots) = 0$

nonlinear self-adjointness exists if there is a substitution $v = \phi(x, u, \ldots)$ such that the adjoint equation $F^*$ is proportional to $F$ on solutions: $$F^*(x, u, v, v_1, \ldots) = \lambda(x, u, \ldots) F(x, u, u_1, \ldots)$$ (Ibragimov, 2011). This property enables systematic construction of conservation laws using symmetry characteristics—via formal Lagrangians—without requiring the system to be of Euler-Lagrange type. The method applies to both scalar and multi-component quasi-linear systems, allowing the construction of local and nonlocal conserved vectors, reduction of order via differential constraints, and explicit identification of physically invariant quantities.

Further, multidimensional integrable systems exhibit an abundance of conservation laws: for pairs (or more) of commuting two-dimensional flows, there exist infinitely many local three-dimensional conservation laws. These are constructed via generating equations dependent on multiple spectral parameters (Makridin et al., 2017). Expansion of these generating functions provides hierarchies of conservation laws for Benney chains, dKP, KdV pairs, and hydrodynamic reductions—a structure central to the Whitham theory and dispersive regularization.

4. Variational and Entropy Methods; Admissibility

Weak solutions for quasi-linear systems require an admissibility criterion. The entropy condition (in the Bardos–Leroux–Nédélec sense) imposes inequalities on weak solutions, selecting the physically relevant ones—crucial in the presence of shocks or discontinuities (Mondal et al., 2016): $$\int_{0}^{T}\int_{\Omega} \operatorname{sgn}(u-k)(f(u) - f(k)) \cdot \nabla \varphi \, dx\,dt + \int_{0}^{T}\int_{\partial\Omega} \operatorname{sgn}(k)(f(y(u)) - f(k))\cdot n \varphi \, dS\,dt \geq 0$$ for all constants $k$ and nonnegative test functions $\varphi$.

Variational methods, such as Hopf–Lax/Lax–Oleinik, extend to non-smooth or non-strictly convex flux functions. The minimizer in the standard representation may fail to be unique; the largest minimizer can be taken to enforce uniqueness in the non-strictly convex (e.g., polygonal flux) case (Caginalp, 2017): $$w(x, t) = g'(y^*(x, t)),\quad y^*(x, t) = \arg^+ \min_{y\in\mathbb{R}}\{tL((x-y)/t) + g(y)\}$$ with $L$ the Legendre transform of the flux.

For systems with boundary conditions, as in open ASEP-type processes, the quasi-potential (large deviation cost function) for scalar conservation laws can be constructed by patching together Riemann problem solutions—optimal paths may involve a combination of rarefaction and shock waves, with minimizers determined via the inverse flux function and Rankine-Hugoniot speeds (Bahadoran, 2010).

5. Integrability, Normal Forms, and Viscous Hierarchies

Perturbative and normal form methods enable classification of (potentially non-Hamiltonian) integrable quasi-linear conservation laws. Deformations parameterized by a viscous central invariant $a(u)$ admit normalization via Miura transformations (Arsie et al., 2013): $$u_t = \partial_x \left(u^2 + \epsilon a(u)u_x + \epsilon^2[(a(u)^2/2)' u_{xx}] + \cdots \right)$$ All integrable deformations with nonzero viscous term are, to leading order, determined by $a(u)$. For $a(u) \equiv 1$, one recovers the classical (viscous) Burgers equation. For $a(u) = u$, one obtains a viscous analog of the Camassa–Holm equation; higher hierarchies map to Burgers and Klein–Gordon hierarchies via hodograph and Miura transformations. Local well-posedness is established for the periodic Cauchy problem (in $H^1$ using Kato’s method), and universality near gradient catastrophe is governed by a viscous analog of the Painlevé I₂ equation.

6. Singularities, Stability, and Numerical Aspects

Discrete schemes for quasi-linear systems introduce artificial viscosity for stability. Spectrally stable shock profiles of the discretization are analyzed via precise asymptotics of the Green’s function associated to the linearized operator (Coeuret, 2023). Stability is deduced under a sharp spectral assumption (Evans function with simple zero at $z = 1$), dispensing with smallness conditions on shock amplitude. The Green’s function decomposes into propagating, reflected, transmitted, and translationally neutral components with heat-kernel type decay estimates. These findings justify numerical approaches in computational fluid dynamics and serve as a foundation for nonlinear orbital stability.

In situations where Riemann invariants degenerate or the hodograph transformation Jacobian vanishes (e.g., in plasticity or gas dynamics with simple waves), conservation laws yield the complete set of solutions as well as characteristics, bridging across singular regions and reconstructing solutions via integral representations (Senashov et al., 2012).

7. Duality, Hamiltonian Structures, and Higher-order Conservation

Systems possessing higher-order Hamiltonian structures (third order Dubrovin–Novikov type) are classified via projective geometry. The corresponding operator is governed by a Monge metric and its factorization (Ferapontov et al., 2017): $$P^{ij} = D_x \left(g^{ij} D_x + c^{ij}_k u_x^k \right) D_x$$ The classification reduces to selecting $n$-tuples of skew-symmetric 2-forms $A^\alpha$ satisfying

$\phi_{\beta\gamma} A^\beta \wedge A^\gamma = 0$

The geometric constraints encode linear degeneracy and Temple class properties: for $n=3$, nonlinearisable systems correspond to WDVV equations; for $n=2$, all systems are linearisable.

Duality for conservation law systems is constructed through ruled hypersurfaces in projective spaces. Hamiltonian systems are autodual, their line generators forming Legendre submanifolds within Fano varieties associated to quadrics (Agafonov, 2019). The geometric approach yields classification of Temple class systems, construction of Abelian relations, and connection to webs of maximal rank.

---

Quasi-linear systems of conservation laws thus constitute a theory at the intersection of nonlinear analysis, geometry, algebra, and physical modeling. Their development continuously informs foundational aspects of mathematical physics, numerical analysis, and the integrable structure of continuum models, guided by symmetry, admissibility principles, and geometric insight.