Numerical-relativity evolution systems for the Einstein equations are deliberately designed so that every characteristic field of the frozen principal part is linearly degenerate in the sense of Lax, on the expectation that genuinely nonlinear gauge shocks then cannot form and smooth initial data evolve smoothly. This record audits the two systems in which the design principle is explicit: the first-order generalized-harmonic system of Lindblom, Scheel, Kidder, Owen and Rinne (2006), which is linearly degenerate exactly for the parameter choice γ1=−1\gamma_1 = -1 (proven by Ji, Lindblom and Cao in 2019), and the Dirac-gauge fully constrained formulation, whose 30-field first-order reduction has all characteristic fields linearly degenerate (Cordero-Carrión et al. 2008) while the well-posedness of the full coupled elliptic-hyperbolic system was explicitly left open by that paper and remains unproven. The literature resolves only the frozen-coefficient algebra: the rigorous shock-prevention theorems (Lax; Liu; Li, Zhou and Kong) are one-dimensional and, for global-in-time statements, small-data; small-data global regularity in wave coordinates is known (Lindblad and Rodnianski); gauge shocks are proven to occur in non-linearly-degenerate gauge sectors (Alcubierre); but no theorem or counterexample addresses the full coupled multidimensional systems. The open problem is to prove or refute, at the natural generality of the full coupled formulations, that linear degeneracy of the principal part excludes Lax-type shock-type gradient catastrophes, and to settle local well-posedness of the coupled fully constrained elliptic-hyperbolic evolution system.

Background

Numerical relativity evolves the Einstein equations as a Cauchy problem, and the choice of coordinates — the gauge — is itself part of the evolution system. A long-standing concern is that the gauge evolution equations can develop gauge shocks: loss of smoothness by crossing characteristics in the coordinate variables (the lapse NN and shift βi\beta^i — in the standard 3+1 decomposition of spacetime into space plus time, the lapse sets how fast proper time advances between successive slices and the shift relabels spatial points from one slice to the next — or the metric components), producing a coordinate singularity in an otherwise regular spacetime. Alcubierre demonstrated such coordinate shocks in hyperbolic formulations of general relativity in 1997, and proved in 2005 that for the Bona-Massó family of hyperbolic slicing conditions (a one-parameter family of algebraic choices of the lapse that includes harmonic slicing) in 1+1 dimensions they are true shocks in the sense of the one-dimensional theory: the characteristic lines of the gauge cross and the foliation loses smoothness. Avoiding gauge shocks is a design requirement for long-term evolutions, in particular with high-order spectral or discontinuous-Galerkin (high-order finite-element) methods in which a discontinuity instantly destroys accuracy.

The mathematical criterion invoked to guarantee shock-free evolution is Lax's dichotomy for first-order quasilinear hyperbolic systems ∂tu+Ak(u) ∂ku=F(u)\partial_t u + A^k(u)\,\partial_k u = F(u). For a characteristic family — an eigenpair (λ,r)(\lambda, r) of the principal symbol nkAkn_k A^k in a spatial direction nn — the family is genuinely nonlinear if the directional derivative of the eigenvalue along the eigenvector does not vanish, and linearly degenerate if

∇uλ⋅r≡0\nabla_u\lambda\cdot r \equiv 0

everywhere in field space. Genuinely nonlinear families steepen smooth data into shocks (finite-time gradient blow-up by characteristic crossing); linearly degenerate families cannot form shocks and admit only contact-type discontinuities. The rigorous theorems behind this dichotomy — singularity formation for genuinely nonlinear one-dimensional systems from smooth compactly supported data (Liu 1979), and the complete global-existence and lifespan theory under (weak) linear degeneracy for small C1C^1 data in one space dimension (Li, Zhou and Kong 1994) — are one-dimensional and, for global-in-time conclusions, small-data. Whether the dichotomy controls genuinely multidimensional coupled nonlinear systems is exactly what is at issue here.

Two production evolution systems were designed around this criterion. First, the generalized-harmonic (GH) formulation fixes coordinates by requiring each coordinate xax^a to satisfy a wave equation of the metric, □gxa=Ha\Box_g x^a = H^a, with a freely chosen gauge-source function HaH_a; the Einstein equations then reduce to a quasilinear wave system for the metric components. Lindblom, Scheel, Kidder, Owen and Rinne (2006) introduced a first-order reduction of this system — the fields (ψab,Πab,Φiab)(\psi_{ab}, \Pi_{ab}, \Phi_{iab}), i.e. the metric, its time derivative and its spatial derivatives — modified by multiples of the constraints, so that the system is symmetric hyperbolic (the principal symbol has a symmetrizer, giving well-posed energy estimates) and, for the parameter choice γ1=−1\gamma_1 = -1, linearly degenerate, while the damping parameters γ0≥0\gamma_0 \ge 0 and γ2>0\gamma_2 > 0 suppress constraint violations. They stated that γ1=−1\gamma_1 = -1 "implies that shocks do not form from smooth initial data", without proof. The algebraic claim was questioned by Cao, Fu, Ji and Xia (2019, online 2018), whose analysis used left eigenvectors (which would additionally require γ2=0\gamma_2 = 0), and then settled by Ji, Lindblom and Cao (2019): with Lax's right-eigenvector definition the system is linearly degenerate if and only if γ1=−1\gamma_1 = -1 — the eigenvalues depend on the fields only through the lapse and shift carried by ψab\psi_{ab}, and the gauge-sector eigenvalue −(1+γ1) nkNk-(1+\gamma_1)\,n_k N^k is independent of the fields exactly at γ1=−1\gamma_1 = -1.

Second, the fully constrained formulation (FCF) of Bonazzola, Gourgoulhon, Grandclément and Novak (2004) evolves the Einstein equations in a 3+1 conformal decomposition γij=ψ4γ~ij\gamma_{ij} = \psi^4\tilde\gamma_{ij} (spatial metric as a conformal factor times a unit-determinant conformal three-metric), enforcing the Hamiltonian and momentum constraints at every time step rather than only at initial time. Maximal slicing (vanishing mean extrinsic curvature, giving an elliptic equation for the lapse NN) and the Dirac gauge Djγ~ij=0\mathcal{D}_j\tilde\gamma^{ij} = 0 (vanishing divergence of the conformal metric with respect to the flat fiducial metric fijf_{ij}) close the system into quasilinear elliptic equations for NN, the conformal factor ψ\psi and the shift βi\beta^i, coupled to a wave-type hyperbolic equation for hij=γ~ij−fijh^{ij} = \tilde\gamma^{ij} - f^{ij}, the deviation of the conformal metric from flat. Cordero-Carrión, Ibáñez, Gourgoulhon, Jaramillo and Novak (2008) analysed the first-order reduction of the hyperbolic sector (30 fields: hijh^{ij}, its time derivative and its spatial derivatives): the system is strongly hyperbolic (the principal symbol has a complete set of eigenvectors) whenever the evolution vector tμt^\mu is timelike, all eigenvalues coincide with the coordinate speed of light or the shift, and all 30 characteristic fields are linearly degenerate — so that "no shocks can be propagated along these curves, in particular gauge shocks", any discontinuities being contact discontinuities. Crucially, this analysis treats NN, ψ\psi and βi\beta^i as frozen parameters; the same paper explicitly leaves the well-posedness of the complete coupled elliptic-hyperbolic system open, and flags that its elliptic sector (closely related to the extended conformal thin-sandwich, XCTS, system used for initial data) can suffer from non-uniqueness. Later work reformulated the elliptic sector to overcome local-uniqueness problems (Cordero-Carrión et al. 2009; Santos-Pérez, Cordero-Carrión and Cerdá-Durán 2025), but no well-posedness theorem for the coupled evolution system exists.

The neighbouring rigorous results sharpen the gap rather than close it. Andersson and Moncrief (2003) proved local strong well-posedness for a coupled elliptic-hyperbolic gauge-fixed Einstein system — constant-mean-curvature slicing plus spatial harmonic coordinates — but in a free-evolution setting where the elliptic equations come only from the gauge, not from elliptically enforced constraints as in the FCF; the 2008 FCF authors explicitly distinguish their constrained case from it. In the small-data regime, Lindblad and Rodnianski (2005) proved global existence for the vacuum Einstein equations in wave coordinates, so small perturbations of Minkowski space in harmonic-type gauges provably never form shocks. What no result covers — for either system — is the full coupled multidimensional nonlinear evolution at large data: whether the linear degeneracy of the frozen principal part actually prevents Lax-type shock formation once the gauge fields, the constraint-damping terms and, for the FCF, the elliptic solves evolve together with the metric.

Problem Statement

The problem is to determine, rigorously and at the natural generality of the full coupled formulations (not frozen-coefficient, constant-gauge or linearized reductions), whether linear degeneracy of the principal part actually guarantees shock-free smooth evolution for the two Einstein-evolution systems in which this design principle is embedded.

(1) Generalized-harmonic system. For the first-order generalized-harmonic vacuum Einstein evolution system of Lindblom, Scheel, Kidder, Owen and Rinne (2006) — including its gauge-source functions and constraint-damping terms, with the linear-degeneracy parameter choice γ1=−1\gamma_1 = -1 (and γ2>0\gamma_2 > 0) — prove or refute that smooth initial data evolve without Lax-type shock formation — that is, that no smooth, constraint-satisfying initial-data set has a classical solution that develops a shock-type gradient catastrophe (characteristic crossing with finite-time blow-up of first derivatives of the gauge or gravitational variables) in finite time.

(2) Fully constrained formulation. For the Dirac-gauge fully constrained formulation, establish local well-posedness — existence, uniqueness, continuous dependence on data, and a continuation criterion — of the complete coupled elliptic-hyperbolic evolution system (elliptic lapse, conformal factor and shift with elliptically enforced Hamiltonian and momentum constraints, coupled to the hyperbolic hijh^{ij} subsystem whose frozen 30-field reduction has all characteristic fields linearly degenerate), or prove a genuine obstruction, for example non-unique evolutions from identical smooth initial data through non-unique elliptic solutions. Establishing (2) includes determining whether the shock-free property suggested by the linear degeneracy of the frozen hyperbolic part survives the elliptic-hyperbolic coupling.

Both deliverables serve one objective: replacing the design principle "linear degeneracy ⇒\Rightarrow no gauge shocks, smooth data stay smooth" by a theorem or a counterexample for the full coupled Einstein-evolution systems. Answers valid only for frozen-coefficient reductions, or only for gauge sources restricted to constants, do not settle the problem.

References

  1. P. D. Lax, "Hyperbolic Systems of Conservation Laws and the Mathematical Theory of Shock Waves", CBMS-NSF Regional Conference Series in Applied Mathematics 11, SIAM, Philadelphia (1973). ISBN 0-89871-177-0. DOI 10.1137/1.9781611970562. Zbl 0268.35062. URL: https://epubs.siam.org/doi/book/10.1137/1.9781611970562
  2. T.-P. Liu, "Development of singularities in the nonlinear waves for quasilinear hyperbolic partial differential equations", Journal of Differential Equations 33, 92–111 (1979). DOI 10.1016/0022-0396(79)90082-2. Zbl 0379.35048. URL: https://doi.org/10.1016/0022-0396(79)90082-2
  3. T.-T. Li, Y. Zhou and D.-X. Kong, "Weak linear degeneracy and global classical solutions for general quasilinear hyperbolic systems", Communications in Partial Differential Equations 19 (7–8), 1263–1317 (1994). DOI 10.1080/03605309408821055. Zbl 0810.35054. URL: https://doi.org/10.1080/03605309408821055
  4. M. Alcubierre, "Appearance of coordinate shocks in hyperbolic formalisms of general relativity", Physical Review D 55, 5981–5991 (1997). DOI 10.1103/PhysRevD.55.5981. URL: https://doi.org/10.1103/PhysRevD.55.5981
  5. M. Alcubierre, "Are gauge shocks really shocks?", Classical and Quantum Gravity 22, 4071–4082 (2005). arXiv:gr-qc/0503030. DOI 10.1088/0264-9381/22/19/017. URL: https://arxiv.org/abs/gr-qc/0503030
  6. B. Reimann, M. Alcubierre, J. A. González and D. Núñez, "Constraint and gauge shocks in one-dimensional numerical relativity", Physical Review D 71, 064021 (2005). arXiv:gr-qc/0411094. DOI 10.1103/PhysRevD.71.064021. URL: https://arxiv.org/abs/gr-qc/0411094
  7. L. Lindblom, M. A. Scheel, L. E. Kidder, R. Owen and O. Rinne, "A New Generalized Harmonic Evolution System", Classical and Quantum Gravity 23, S447–S462 (2006). arXiv:gr-qc/0512093. DOI 10.1088/0264-9381/23/16/S09. URL: https://arxiv.org/abs/gr-qc/0512093
  8. L. Andersson and V. Moncrief, "Elliptic-Hyperbolic Systems and the Einstein Equations", Annales Henri Poincaré 4, 1–34 (2003). arXiv:gr-qc/0110111. DOI 10.1007/s00023-003-0120-1. URL: https://arxiv.org/abs/gr-qc/0110111
  9. H. Lindblad and I. Rodnianski, "Global Existence for the Einstein Vacuum Equations in Wave Coordinates", Communications in Mathematical Physics 256, 43–110 (2005). DOI 10.1007/s00220-004-1281-6. URL: https://doi.org/10.1007/s00220-004-1281-6
  10. S. Bonazzola, E. Gourgoulhon, P. Grandclément and J. Novak, "Constrained scheme for the Einstein equations based on the Dirac gauge and spherical coordinates", Physical Review D 70, 104007 (2004). DOI 10.1103/PhysRevD.70.104007. URL: https://doi.org/10.1103/PhysRevD.70.104007
  11. I. Cordero-Carrión, J. M. Ibáñez, E. Gourgoulhon, J. L. Jaramillo and J. Novak, "Mathematical Issues in a Fully-Constrained Formulation of Einstein Equations", Physical Review D 77, 084007 (2008). arXiv:0802.3018. DOI 10.1103/PhysRevD.77.084007. URL: https://arxiv.org/abs/0802.3018
  12. I. Cordero-Carrión, P. Cerdá-Durán, H. Dimmelmeier, J. L. Jaramillo, J. Novak and E. Gourgoulhon, "Improved constrained scheme for the Einstein equations: An approach to the uniqueness issue", Physical Review D 79, 024017 (2009). arXiv:0809.2325. DOI 10.1103/PhysRevD.79.024017. URL: https://arxiv.org/abs/0809.2325
  13. I. Cordero-Carrión, P. Cerdá-Durán and J. M. Ibáñez, "Gravitational waves in dynamical spacetimes with matter content in the fully constrained formulation", Physical Review D 85, 044023 (2012). arXiv:1108.0571. DOI 10.1103/PhysRevD.85.044023. URL: https://arxiv.org/abs/1108.0571
  14. Z. Cao, P. Fu, L.-W. Ji and Y. Xia, "Application of local discontinuous Galerkin method to Einstein equations" (title of the published version; the arXiv preprint is Part II of "Binary black hole simulation with an adaptive finite element method"), International Journal of Modern Physics D 28, 1950014 (2019). arXiv:1805.10640. DOI 10.1142/S0218271819500147. URL: https://arxiv.org/abs/1805.10640
  15. L.-W. Ji, L. Lindblom and Z. Cao, "Linear Degeneracy of the First-Order Generalized-Harmonic Einstein System", Classical and Quantum Gravity 36, 087001 (2019). arXiv:1903.00140. DOI 10.1088/1361-6382/ab0d36. URL: https://arxiv.org/abs/1903.00140
  16. T. W. Baumgarte and D. Hilditch, "Shock-avoiding slicing conditions: tests and calibrations", Physical Review D 106, 044014 (2022). arXiv:2207.06376. DOI 10.1103/PhysRevD.106.044014. URL: https://arxiv.org/abs/2207.06376
  17. M. Dumbser, F. Guercilena, S. Köppel, L. Rezzolla and O. Zanotti, "Conformal and covariant Z4 formulation of the Einstein equations: strongly hyperbolic first-order reduction and solution with discontinuous Galerkin schemes", Physical Review D 97, 084053 (2018). arXiv:1707.09910. DOI 10.1103/PhysRevD.97.084053. URL: https://arxiv.org/abs/1707.09910
  18. S. Santos-Pérez, I. Cordero-Carrión and P. Cerdá-Durán, "Reformulation of Einstein equations in the fully constrained formulation: Local-uniqueness, post-Newtonian expansion, and initial data", Physical Review D 112, 024057 (2025). arXiv:2503.10323. DOI 10.1103/tbz7-hpcy. URL: https://arxiv.org/abs/2503.10323