Characteristic, Hyperboloidal, and Boundary Problems

Einstein evolution changes character when data are placed on null hypersurfaces, on spacelike slices reaching null infinity, or beside a timelike boundary. The same field equations then divide into transport constraints, hyperbolic evolution, boundary fluxes, and corner compatibility conditions in different ways. This chapter develops those divisions without treating a coordinate boundary, null infinity, and timelike conformal infinity as interchangeable objects.

Helpful background. The Einstein Equation supplies the Bianchi identity; Analysis for Mathematical Relativity supplies regularity and energy estimates; Hypersurface Normals and Tangential Projectors supplies induced geometry; PDEs and Green Functions for Gravitation reviews characteristics; Causality and Global Structure clarifies domains of dependence and conformal boundaries.

Enter the characteristic and boundary problems

The ordinary Cauchy problem uses a spacelike hypersurface, whose future unit normal is transverse to the surface. A null hypersurface is different: its normal is tangent, and the induced metric is degenerate. Some Einstein equations therefore become ordinary differential equations along its null generators. A timelike hypersurface presents the opposite issue: causal signals can enter through it, so incoming fields require boundary data.

Hyperboloidal slices are spacelike and determine an evolution in their domain of dependence while approaching future null infinity. They need not be Cauchy surfaces for an entire ambient spacetime: a future Minkowski hyperboloid, for example, covers a different causal region from a complete inertial plane. A conformal compactification can bring null infinity to a finite coordinate location, provided the rescaled data meet strong regularity conditions.

Three distinctions recur throughout the chapter:

• free data versus constrained data: characteristic and boundary fields cannot all be chosen independently;
• gauge uniqueness versus geometric uniqueness: a reduced hyperbolic system is unique in its gauge, while the spacetime is unique only up to a data-fixing diffeomorphism;
• local existence versus global behavior: a regular corner theorem does not prevent later caustics, curvature concentration, or boundary-driven instability.

The six-page route

The pages are ordered so that every new kind of boundary is introduced only after its local data have been understood.

1. Existence Theory for the Characteristic Initial-Value Problem states the smooth local theorem for two intersecting null hypersurfaces. It identifies the null constraints, admissible free data, corner compatibility, and geometric uniqueness.
2. Double-Null Evolution introduces optical functions, null frames, expansions, shears, curvature components, and the transport–elliptic hierarchy used in estimates.
3. Hyperboloidal Data and Conformal Infinity compactifies a spacelike slice reaching $\mathscr I^+$ and explains shear-free and higher regularity conditions.
4. Well-Posedness of the Einstein Initial-Boundary Value Problem treats a finite timelike boundary through symmetric-hyperbolic reduction, maximally dissipative data, subsidiary constraints, and corner jets.
5. Constraint-Preserving and Radiation-Controlling Boundaries separates incoming physical, constraint, and gauge sectors and shows why each must be handled.
6. Corner Compatibility and Anti-de Sitter Boundaries compares finite-boundary, asymptotically AdS, unique-continuation, and instability results through the evidence cutoff 2026-08-12.

Two ways through the chapter

For characteristic geometry, read pages 1–3. This route begins with local existence, exposes the double-null equations that make it work, and ends at null infinity. It prepares the later chapters on trapped surfaces, black-hole decay, and radiation theorems.

For timelike boundaries, read pages 4–6, consulting page 2 for null characteristic fields. This route begins with an abstract well-posedness theorem, separates constraint and radiation control, and then asks what changes when the boundary is conformal infinity in AdS.

Readers interested in proofs should track the same pattern in every formulation:

$$\begin{aligned} \text{reduced hyperbolic system} &\longrightarrow\text{energy estimate}\\ &\longrightarrow\text{subsidiary propagation}\\ &\longrightarrow\text{geometric solution}. \end{aligned}$$

The data and boundary operators change, but omitting the subsidiary step always leaves open whether a reduced solution actually satisfies Einstein’s equations.

The smooth intersecting-null theorem and its reduction to a spacelike problem are proved in Rendall 1990, pp. 221–239; this exact scope is the reference point for the chapter’s later extensions.

For timelike boundaries, compare the finite-boundary theorem in Friedrich–Nagy 1999, Theorem 8.1 with the prescribed conformal-infinity problem in Enciso–Kamran 2019, Theorem 1.1. Their hypotheses and data are different; the final page compares them explicitly.

Convention-sensitive points

The null pages use $g(e_3,e_4)=-2$ when writing expansions and curvature components. Rescaling the null pair changes individual expansions. The hyperboloidal pages use $\widetilde g=\Omega^2g$ for the unphysical metric. On spacelike slices, whenever extrinsic curvature appears it uses

$$K_{ij}=-h_i{}^\mu h_j{}^\nu\nabla_\mu n_\nu.$$

Boundary flux signs depend on the outward normal. At the inner edge of a half-space, the outward normal can point opposite to the increasing coordinate direction; this is a common source of an incorrect dissipativity sign.

Review the chapter

1. Classify the data surface. For each of the following, say whether arbitrary incoming data are required: a spacelike Cauchy surface without boundary, one null hypersurface, two intersecting null hypersurfaces, future null infinity, and a timelike AdS boundary.

Solution

A boundary-free spacelike Cauchy surface carries constrained initial data but no later incoming boundary data. One null hypersurface normally needs a second null branch, a worldtube, or vertex regularity information. Two intersecting null hypersurfaces enclose a local domain once their constrained data agree at the corner. Future null infinity is characteristic, so arbitrary incoming physical data are not prescribed there. Timelike AdS infinity can send signals into the bulk, so a boundary condition is necessary.

2. Diagnose a failed argument. A reduced Einstein IBVP has a nonincreasing main energy, and its initial constraints vanish. Why does this not yet prove that the evolved metric satisfies Einstein’s equations?

Solution

The main energy estimate establishes well posedness only for the reduced variables. On a region with timelike boundary, the homogeneous subsidiary constraint equations also have incoming characteristic fields. Unless the main boundary conditions force homogeneous boundary data for that subsidiary system, constraint violations can enter later despite vanishing initial constraints.

3. Transfer task. Explain why local smooth AdS existence and evidence for weakly turbulent collapse are logically compatible.

Solution

Local existence controls a finite interval whose length depends on the data norms and boundary conditions. Turbulent-collapse claims concern long-time transfer of energy across scales under particular reflecting conditions. A solution can satisfy the local theorem and later concentrate. The statements also differ in symmetry, matter model, and proof status, so neither negates the other.

The next chapter asks what remains meaningful when the metric or its derivatives are too rough for these smooth arguments: Low-Regularity Spacetimes and Weak Solutions.

References

• Enciso, Alberto, and Niky Kamran. “Lorentzian Einstein Metrics with Prescribed Conformal Infinity.” Journal of Differential Geometry 112 (2019): 505–554. DOI. Open PDF.
• Friedrich, Helmut, and Gabriel Nagy. “The Initial Boundary Value Problem for Einstein’s Vacuum Field Equation.” Communications in Mathematical Physics 201 (1999): 619–655. Open PDF.
• Rendall, Alan D. “Reduction of the Characteristic Initial Value Problem to the Cauchy Problem and Its Applications to the Einstein Equations.” Proceedings of the Royal Society A 427 (1990): 221–239. DOI.