1: \input amstex
2:
3:
4: \magnification=1150
5:
6:
7:
8: \centerline{\bf A Combinatorial Curvature Flow for Compact 3-Manifolds with Boundary}
9:
10: \medskip
11: \centerline{\bf Feng Luo}
12: \medskip
13: \centerline{\bf abstract}
14: \medskip
15: We introduce a combinatorial curvature flow for PL metrics on compact triangulated 3-manifolds with boundary consisting of surfaces
16: of negative Euler characteristic. The flow tends to find the complete hyperbolic metric with totally geodesic boundary on a
17: manifold. Some of the basic properties of the combinatorial flow are established.
18: The most important ones is that the evolution of the combinatorial curvature satisfies a combinatorial heat
19: equation. It implies that the total curvature decreases along the flow. The local convergence of the flow to
20: the hyperbolic metric is also established if the triangulation is isotopic to a totally
21: geodesic triangulation.
22:
23:
24: \medskip
25: \noindent
26: {\it Mathematics Subject Classification: 53C44, 52A55}
27:
28: \medskip
29: \noindent
30: \S1. {\bf Introduction}
31:
32: \medskip
33: \noindent
34: 1.1. The purpose of this paper is to announce the construction of a combinatorial curvature flow which is a 3-dimensional
35: analogy of the work of [CL]. In [CL], we introduced a 2-dimensional combinatorial curvature flow for triangulated
36: surfaces of non-positive Euler characteristic. It is shown that for any initial choice
37: of PL metric of circle packing type, the flow exists for all time and converges exponentially fast to the
38: Andreev-Koeb-Thurston's circle packing metrics.
39:
40: The basic building blocks for the 2-dimensional flow are hyperbolic and Euclidean triangles.
41: In the 3-dimensional case, the basic building blocks are the \it hyperideal tetrahedra \rm discovered by
42: Bao and Bonahon [BB]. Given an ideal triangulation of a compact 3-manifold with boundary consisting of
43: surfaces of negative Euler characteristic, we replace each (truncated) tetrahedron by a hyperideal tetrahedron by
44: assigning the edge lengths. The isometric gluing of these tetrahedra gives a hyperbolic cone metric on the 3-manifold.
45: The \it PL curvature \rm of the cone metric at an edge is $2\pi$ less than the sum of dihedral angles at the edge.
46: The combinatorial curvature flow that we propose is the following system of ordinary differential equations,
47: $$ d x_i/dt = K_i \tag 1.1$$
48: where $x_i$ is the length of the i-th edge and $K_i$ is the curvature of the cone metric $(x_1, ..., x_n)$ at the i-th edge.
49: The equation (1.1) captures the essential features of the 2-dimensional combinatorial Ricci flow in [CL].
50: The most important of all is that the PL curvature evolves according to a combinatorial heat equation. Thus the corresponding
51: maximal principle applies. The flow has
52: the tendency of finding the complete hyperbolic metric of totally geodesic boundary on the manifold.
53: By analyzing the singularity formations in equation (1.1), it is conceivable that one could give a new proof of
54: Thurston's geometrization theorem for these manifolds using (1.1). Furthermore, the flow (1.1) will be a useful tool to find algorithmically
55: the complete hyperbolic metric.
56:
57: \medskip
58: \noindent
59: 1.2. We now provide some details of the approach. Suppose $M$ is a compact 3-manifold with
60: boundary consisting of surfaces of negative Euler characteristic. An \it ideal triangulation \rm (or truncated triangulation)
61: of $M$ is
62: the following finite set of data. Take a finite collection of 3-simplexes
63: and identify their faces in pairs by homeomorphisms.
64: The quotient space with a small regular neighborhood of each vertex removed
65: is homeomorphic to the 3-manifold $M$. By abuse the use of language, we will call (the homeomorphism images of)
66: i-dimensional cells ($i \geq 1$) of
67: the ideal triangulation (truncated triangulation) in the interior of $M$, the \it edges, triangles and tetrahedra. \rm
68: A \it hyperideal tetrahedron \rm in the 3-dimensional hyperbolic space is a compact convex polyhedron so that
69: it is diffeomorphic to a truncated tetrahedron in the 3-dimensional Euclidean space and its four hexagonal
70: faces are right-angled hyperbolic hexagons. (Note that
71: two compact subsets of $\bold R^n$ are diffeomorphic if there is a diffeomorphism between two open neighborhoods of
72: them sending one compact set to the other). See the figure below. In the beautiful work of [BB], Bao and Bonahon give a complete characterization
73: of hyperideal convex polyhedra. As a consequence of their work, hyperideal
74: tetrahedra are completely characterized by their six dihedral angles at the six edges so that the sum of three dihedral angles associated
75: to edges adjacent to each vertex is less than $\pi$.
76:
77: \medskip
78:
79: \input epsf
80: \centerline{\epsfbox{p.eps}}
81: \centerline{\it The six edges of a hyperideal tetrahedron are the intersections of its hexagonal faces \rm}
82: \centerline{\bf figure }
83:
84:
85: \medskip
86: One of our main technical observations is the following,
87:
88: \medskip
89: \noindent
90: {\bf Theorem 1.} \it The volume of a hyperideal tetrahedron is a strictly
91: concave function of its dihedral angles. \rm
92: \medskip
93:
94: Note that each hyperideal tetrahedron is also determined by its six edge lengths. By the Schlaefli formula, an equivalent statement to
95: theorem 1 is that the Hessian matrix of the map sending six edge lengths to the six dihedral angle is strictly positive definite.
96: This positive definite matrix provides a basis for constructing the combinatorial Laplacian operator for the curvature evolution equation.
97:
98:
99: \medskip
100: Now given an ideal triangulated 3-manifold $(M, T)$, let $E$ be the set of edges in the triangulation and let $n$ be the number
101: of edges in $E$. An assignment $x: E \to \bold R_{>0}$ is called a \it hyperbolic cone metric associated to the triangulation
102: $T$ \rm if for each tetrahedron $t$ in $T$ with edges $e_1, ..., e_6$, the six numbers $x_i = x(e_i)$ (i=1,..., 6) are the
103: edge lengths of a hyperideal tetrahedron in $\bold H^3$. The set of all hyperbolic cone metrics associated to $T$ is
104: denoted by $L(M, T)$ which will be considered as an open subset of $\bold R^n = \bold R^E$. The PL curvature of a cone metric
105: $x \in L(M,T)$ is the map $K(x): E \to \bold R$ sending an edge $e$ to the PL curvature of $x$ at the edge $e$. Again we
106: identify the
107: set of all PL curvatures $\{ K(x) | x \in L(M, T)\}$ with a subset of $\bold R^n$. The
108: combinatorial curvature flow is the vector field in $L(M, T)$ defined by equation (1.1) where $K_i$ in the
109: right-hand side is the PL curvature of the metric $x=(x_1,...,x_n)$ at time $t$ at the i-th edge.
110:
111:
112: \medskip
113: \noindent
114: {\bf Theorem 2.} \it For any ideal triangulated 3-manifold, under the combinatorial curvature flow (1.1), the PL curvature
115: $K_i(t)$ evolves according to a combinatorial heat
116: equation,
117: $$ dK_i(t)/dt = \sum_{j=1}^n a_{ij} K_j(t) \tag 1.2$$
118: where the matrix $[a_{ij}]_{n \times n}$ is symmetric negative definite.
119: In particular, the total curvature
120: $\sum_{i=1}^n K_i^2(t)$ is strictly decreasing along the flow unless $K_i(t) =0$ for all $i$.
121: \rm
122:
123: \medskip
124: \noindent
125: {\bf Theorem 3.} \it For any ideal triangulated 3-manifold $(M,T)$,
126: the equilibrium points of the combinatorial curvature flow (1.1) are the complete hyperbolic
127: metric with totally geodesic boundary. Furthermore, each equilibrium point is a local attractor of the flow.\rm
128:
129: \medskip
130:
131: Another consequence of the convexity is the following local rigidity result for hyperbolic cone metrics without constrains on
132: cone angles. Note that by [HK], hyperbolic cone metrics with cone angles at most $2\pi$ are locally rigid.
133:
134: \medskip
135: \noindent
136: {\bf Theorem 4. } \it For any ideal triangulated 3-manifold $(M,T)$, the curvature map $\Pi: L(M, T) \to \bold R^n$ sending
137: a metric $x$ to its curvature $K(x)$ is a local diffeomorphism. In particular, a hyperbolic cone metric associated to
138: an ideal triangulation is locally determined by its cone angles. \rm
139:
140:
141:
142: \medskip
143: \noindent
144:
145: In the rest of the note, we will sketch briefly the ideas of the proofs of the results. In the last section, we propose several
146: questions related to the combinatorial curvature flow whose resolution will lead to a new proof of Thurston's geometrization theorem
147: for this class of 3-manifolds.
148:
149: \medskip
150: \noindent
151: 1.3. Acknowledgment. We thank Ben Chow for conversations on the topic of Ricci flow and X.S. Lin for
152: discussions. The work is supported in part by the NSF and a research grant from the Rutgers University.
153:
154: \medskip
155: \noindent
156: \S2. {\bf Sketch of the Proofs}
157: \medskip
158: \noindent
159: 2.1. The key step in the proof of all theorems is to establish the convexity theorem 1. Here is a quick way to see it.
160: Suppose $x_1, ..., x_6$ are the lengths of the six edges of a hyperideal tetrahedron so that
161: the corresponding dihedral angles are $a_1, ..., a_6$. Let $a=(a_1, ..., a_6)$, $x=(x_1, ..., x_6)$ and $V$ be the
162: volume of the hyperideal tetrahedron. Then
163: by using cosine laws for hyperbolic triangles and hyperbolic right-angled hexagons, we see that both
164: functions $x=x(a)$ and $a=a(x)$ are smooth in $a$ and $x$ respectively. In particular, they are diffeomorphisms. By the Schlaefli formula $\partial V/\partial a_i = -x_i/2$,
165: the Jacobian matrix $[\partial x_i /\partial a_j]_{6 \times 6}$ is
166: symmetric. The Jacobian matrix is non-singular due to the fact that the map $a=a(x)$ is a diffeomorphism. Thus its signature is
167: independent of the choice of hyperideal tetrahedra. One checks directly that in the case the hyperideal tetrahedron is regular (i.e., all $x_i$'s are the same and
168: all $a_i$'s are the same), the Jacobian matrix is positive definite. Thus theorem 1 follows. Since the inverse of
169: a symmetric positive definite matrix is again symmetric and positive definite, we see that the matrix
170: $[\partial a_i/\partial x_j]_{6 \times 6}$ is also symmetric and positive definite.
171:
172: \medskip
173: \noindent
174: 2.2. To prove theorem 2, one uses the equation (1.1). Namely, to understand the evolution of the curvature, it suffices to
175: understand the evolution of the individual dihedral angle $a_i(t)$. By the chain rule, we have,
176: $$ da_i/dt = \sum_{j=1}^6 (\partial a_i/\partial x_j) dx_j/dt = \sum_{j=1}^6 (\partial a_i/\partial x_j) K_j. \tag 2.1$$
177: By the discussion above, the matrix $[\partial a_i/\partial x_j]_{6 \times 6}$ is positive definite.
178: By summing over all dihedral angles adjacent to a fixed edge, we obtain theorem 2 since the sum of semi-negative definite matrices
179: are still semi-negative definite. A further study shows that the resulting matrix is strictly negative definite.
180:
181: \medskip
182: \noindent
183: 2.3. Both theorems 3 and 4 follow from the fact that the combinatorial curvature flow (1.1) is the negative gradient flow of a
184: strictly convex function defined on the space $L(M, T)$. Indeed, by the Schlaefli formula $dV = -1/2 \sum_i x_i da_i$,
185: if we form $H = 2V +\sum_{i=1}^6 a_i x_i$, then
186: $\partial H/\partial x_i = a_i$. Thus the function $H$ is a strictly convex function of the edge lengths $x_1, ..., x_6$.
187: For any hyperbolic cone metric $x$ associated to the ideal triangulation $T$, we define
188: $$H(x) = 2 vol(x) - \sum_{i} K_i x_i \tag 2.2$$
189: where $vol(x)$ is the volume of the metric. One checks easily that $H(x)$ is a strictly convex function of $x$ and
190: the gradient of $H(x)$ is $-(K_1, ..., K_n)$. Thus theorem 3 follows from the standard theory of ordinary differential equation. To see theorem 4, one uses the fact that the map sending a point to the gradient of a strictly smooth convex function is a local diffeomorphism.
191:
192: \medskip
193: \noindent
194: \S 3. {\bf Some Remarks and Questions}
195: \medskip
196: This work and [CL] are motivated by the work of Richard Hamilton [Hi] on Ricci flow. The strategy of
197: [Hi] is to find a flow deforming metrics so that its curvature evolves according to a heat-type equation. The main focus of
198: study then shifts from the evolution of the metrics to the evolution of its curvature using the maximal principle for heat equations.
199: As long as one has control of the curvature evolution, then one gets some control of the metric evolutions by study either the
200: singularity formation or the long time convergence. Theorem 2 above seems to indicate that the combinatorial flow (1.1) deforms
201: the cone metrics in the "right" direction. There remains the task of understanding the singularity formations in (1.1) which corresponds
202: to the degeneration of the hyperideal tetrahedra. This is being investigated. Below are some thoughts on this topic.
203: The motivations come from the work of [Th], [CV], [Riv2], [Lei] and [CL].
204:
205: \medskip
206: \noindent
207: 3.1. To understand the singularity formation, we will focus our attention to the function $H$ in (2.2).
208: Following [CV] and [Riv2], our goal is to find the (linear) conditions on the ideal triangulation so that it will guarantee the
209: existence of critical points for $H$. The existence of the ideal triangulation satisfying the (linear) condition will be
210: related to the topology of the 3-manifold and will be resolved by topological arguments.
211:
212: Suppose $(M, T)$ is an ideal triangulated compact 3-manifold so that each boundary component
213: of $M$ has negative Euler characteristic. A pair $(e, t)$ where $e$ is an edge and $t$ is a tetrahedron containing $e$ is
214: called a \it corner \rm in $T$. Following Rivin [Riv1], Casson and Lackenby [La1], we say the triangulated manifold $(M, T)$
215: supports a \it linear hyperbolic structure \rm if one can assign
216: each corner of $T$ a positive number called the \it dihedral angle \rm so that (1) the sum of dihedral angles
217: of all corners adjacent to each fixed edge
218: is $2\pi$, and (2) the sum of dihedral angles of every triple of corners $(e_1, t), (e_2, t), (e_3, t)$ where $e_1, e_2, e_3$
219: are adjacent to a fixed vertex is less than $\pi$. By the work of [BB], given a linear hyperbolic structure on $(M, T)$, we can realize
220: each individual tetrahedron by a hyperideal tetrahedron whose dihedral angles are the assinged numbers so that the sum of
221: the dihedral angles at each edge is $2\pi$.
222: It can be shown that if $(M, T)$ supports a linear hyperbolic structure,
223: then the manifold $M$ is irreducible without incompressible tori. One would ask if the converse is also true. The work of Lackenby [La2] gives
224: some evidences that the following may have a positive answer. See also the work of [KR].
225:
226: \medskip
227: \noindent
228: {\bf Question 1.} \it Suppose $M$ is a compact irreducible 3-manifold with incompressible boundary consisting of surfaces
229: of negative Euler characteristic. If $M$ contains no incompressible tori and annuli, is there any ideal triangulation of $M$ which supports a
230: linear hyperbolic structure? \rm
231:
232: \medskip
233: The next question is the 3-dimensional analogy of the 2-dimensional singularity formation analysis in the work of [CV], [Riv2]
234: and [Le].
235: \medskip
236: \noindent
237: {\bf Question 2.} \it Suppose $(M, T)$ is an ideal triangulated 3-manifold which supports a linear hyperbolic structure.
238: Does $H$ have a local minimal point in the space $L(M,T)$ of all cone metrics associated to $(M, T)$?
239: \rm
240:
241:
242: \medskip
243: Positive resolutions of above two questions will produce a new proof of Thurston's geometrization theorem for this
244: class of 3-manifolds.
245:
246: %\medskip
247: %\noindent
248: %3.2. The convexity of the volume function in terms of the dihedral angles has been the key to the work of Rivin [Riv2] and
249: %Leibon [Lei]. By looking at these results, one is attemped to ask if there is any general result concerning the
250: %strictly convexity of the volume function in terms of the dihedral angles.
251:
252: \medskip
253: \noindent
254: 3.2. Suppose $(M,T)$ supports a linear hyperbolic structure. We define the
255: volume of a linear hyperbolic structure to be the sum of the volumes of
256: its hyperideal tetrahedra.
257: Let
258: $LH(M,T)$ be the space of all linear hyperbolic structures on $(M,T)$.
259: It can be shown, using Lagrangian multipliers, that the volume function is
260: strictly concave on $LH(M,T)$ whose maximal point is exactly the complete
261: hyperbolic metric on $M$. The situation is the same as ideal triangulations
262: of compact 3-manifolds whose boundary consists of tori. In this case, one realizes each tetrahedron by an ideal tetrahedron in the hyperbolic space. It can
263: be shown that
264: the complete hyperbolic metric of finite volume is exactly equal to the
265: maximal point of the volume function defined on the space of all linear hyperbolic structures defined in [Ri1]. This was also observed by Rivin [Ri3].
266: \medskip
267: \noindent
268: 3.3. We remark that the moduli space of all hyperideal tetrahedra parametrized by their edge lengths $x_1, ..., x_6$ is not
269: a convex subset of $\bold R^6$. This is the main reason that we have only local convergence and local rigidity in theorems
270: 3 and 4.
271: However, it is conceivable that
272: theorem 4 may still be true globally. We do not know yet if the
273: space $L(M, T)$ of all cone metrics associated to the ideal triangulated manifold is homeomorphic to a Euclidean space.
274: It is likely to be the case. In fact, one would hope that there is a diffeomorphism $h: \bold R_{>0} \to \bold R_{>0}$ so
275: that if we parameterize the space of all hyperideal tetrahedra by $(t_1, ..., t_6)=(h(x_1), ..., h(x_6))$, then
276: the space becomes convex in $t$-coordinate. Evidently, if this holds, it implies the space $L(M, T)$ is convex
277: in the t-coordinate.
278:
279: \medskip
280: \medskip
281:
282: \centerline{\bf References}
283: \medskip
284:
285: \noindent
286: [BB] Bao, Xiliang; Bonahon, Francis: Hyperideal polyhedra in hyperbolic 3-space. Bull. Soc. Math. France 130 (2002), no. 3, 457--491.
287:
288: \noindent
289: [CL] Chow, Bennett; Luo, Feng: Combinatorial Ricci flows on surfaces. J. Differential Geom. 63 (2003), no. 1, 97--129.
290:
291: \noindent
292: [CV] Colin de Verdière, Yves: Un principe variationnel pour les empilements de cercles. Invent. Math. 104 (1991), no. 3, 655--669.
293:
294: \noindent
295: [Hi] Hamilton, Richard S. Three-manifolds with positive Ricci curvature. J. Differential Geom. 17 (1982), no. 2, 255--306.
296:
297: \noindent
298: [HK] Hodgson, Craig D.; Kerckhoff, Steven P: Rigidity of hyperbolic cone-manifolds and hyperbolic Dehn surgery. J. Differential Geom. 48 (1998), no. 1, 1--59.
299:
300: \noindent
301: [KR] Kang, Ensil and JH Rubinstein, J. H: ideal triangulations of 3-manifolds I, preprint.
302:
303: \noindent
304: [La1] Lackenby, Marc: Word hyperbolic Dehn surgery. Invent. Math. 140 (2000), no. 2, 243--282.
305:
306: \noindent
307: [La2] Lackenby, Marc: Taut ideal triangulations of 3-manifolds. Geom. Topol. 4 (2000), 369--395 (electronic).
308:
309: \noindent
310: [Le] Leibon, Gregory: Characterizing the Delaunay decompositions of compact hyperbolic surfaces. Geom. Topol. 6 (2002), 361--391 (electronic).
311:
312: \noindent
313: [Ri1] Rivin, Igor: Combinatorial optimization in geometry. Adv. in Appl. Math. 31 (2003), no. 1, 242--271.
314:
315: \noindent
316: [Ri2] Rivin, Igor: Euclidean structures on simplicial surfaces and hyperbolic volume. Ann. of Math. (2) 139 (1994), no. 3, 553--580.
317:
318: \noindent
319: [Ri3] Rivin, Igor: private communication.
320:
321: \noindent
322: [Th] Thurston, William: Topology and geometry of 3-manifolds, Lecture notes, Princeton University, 1978.
323: \medskip
324:
325:
326: Department of Mathematics, Rutgers University, Piscataway, NJ 07059
327:
328: Email: fluo\@math.rutgers.edu
329:
330:
331: \end
332: