Geometrization conjecture

At least 49 years old · documented by

A 33-manifold is closed if it is compact and has empty boundary. A closed 33-manifold MM is prime if whenever MM is homeomorphic to a connected sum M1#M2M_1 \# M_2, one of M1,M2M_1, M_2 is homeomorphic to S3S^3.

A model geometry is a pair (X,G)(X,G) consisting of a simply connected smooth manifold XX together with a smooth transitive action of a Lie group GG on XX whose point stabilizers are compact; it is maximal if GG is maximal among Lie groups acting smoothly and transitively on XX with compact stabilizers. A geometric structure on a manifold NN modelled on (X,G)(X,G) is a diffeomorphism from NN to a quotient X/ΓX/\Gamma, where Γ≤G\Gamma \le G is a discrete subgroup acting freely on XX. Fixing a GG-invariant Riemannian metric on XX, such a structure has finite volume if the induced metric on NN gives NN finite total volume.

Up to equivalence there are exactly eight maximal 33-dimensional model geometries admitting at least one compact quotient manifold, namely those with XX equal to

S3,E3,H3,S2×R,H2×R,SL~(2,R),Nil,Sol,S^3,\quad \mathbb{E}^3,\quad \mathbb{H}^3,\quad S^2 \times \mathbf{R},\quad \mathbb{H}^2 \times \mathbf{R},\quad \widetilde{\mathrm{SL}}(2,\mathbf{R}),\quad \mathrm{Nil},\quad \mathrm{Sol},

with GG in each case the full group of isometries of the corresponding invariant metric: O(4,R)O(4,\mathbf{R}) for S3S^3; R3⋊O(3,R)\mathbf{R}^3 \rtimes O(3,\mathbf{R}) for E3\mathbb{E}^3; O+(1,3,R)O^{+}(1,3,\mathbf{R}) for H3\mathbb{H}^3; O(3,R)×R×Z/2ZO(3,\mathbf{R}) \times \mathbf{R} \times \mathbf{Z}/2\mathbf{Z} for S2×RS^2 \times \mathbf{R}; O+(1,2,R)×R×Z/2ZO^{+}(1,2,\mathbf{R}) \times \mathbf{R} \times \mathbf{Z}/2\mathbf{Z} for H2×R\mathbb{H}^2 \times \mathbf{R}; a group with two components and identity component (R×SL~2(R))/Z(\mathbf{R} \times \widetilde{\mathrm{SL}}_2(\mathbf{R}))/\mathbf{Z} and point stabilizer O(2,R)O(2,\mathbf{R}) for SL~(2,R)\widetilde{\mathrm{SL}}(2,\mathbf{R}); the semidirect product of the 33-dimensional Heisenberg group with O(2,R)O(2,\mathbf{R}) for Nil\mathrm{Nil}; and the group of maps of 22-dimensional Minkowski space to itself that are isometries or multiply the metric by −1-1 for Sol\mathrm{Sol}.

Then: for every oriented prime closed 33-manifold MM there is a finite collection T⊂MT \subset M of disjoint embedded incompressible 22-tori such that each connected component of M∖TM \setminus T has interior admitting a finite-volume geometric structure modelled on one of the eight geometries above.

References

Primary source

Wikipedia

Additional references

  1. Wikipedia, Geometrization conjecture, the article this problem comes from.

Progress summary

Refreshed
Claimed solved

The conjecture is a theorem: Perelman’s work completed Thurston’s program for decomposing every suitable three-dimensional space into standard geometric pieces.

William Thurston formulated the conjecture in the late 1970s and published it in 1982. It asserts that every oriented prime closed 33-manifold can be cut along incompressible tori into pieces carrying one of eight standard geometries.

Known results

  • Thurston proved geometrization for Haken manifolds and other substantial classes (1980s).
  • Thurston’s hyperbolization theorem supplied a key ingredient in the later proof.
  • Agol’s 2012 work on virtual Haken and virtual fibering conjectures was related subsequent progress, not a replacement for geometrization.

2002–2003 completion

Grigori Perelman proved the conjecture in full generality using Ricci flow with surgery. Subsequent expositions and references treat the result as established, including the stated decomposition into finite-volume geometric pieces.

Current status (as of August 2026): The geometrization conjecture is settled in full generality by Perelman’s work; no unresolved mathematical case or credible competing claim is reported.

Sources

Solutions 0

No solutions have been posted yet.