14.02.2013 Views

Gruber P. Convex and Discrete Geometry

Gruber P. Convex and Discrete Geometry

Gruber P. Convex and Discrete Geometry

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

198 <strong>Convex</strong> Bodies<br />

Suppose now that any two points of M can be connected by a continuous curve of<br />

finite length. Then the intrinsic or geodesic metric ϱ of the metric space 〈M,δ〉 is<br />

defined as follows:<br />

ϱ(x, y) is the infimum of the lengths of the continuous curves<br />

in M connecting x, y for any x, y ∈ M.<br />

In general the metrics δ <strong>and</strong> ϱ will be different, but they induce the same topology on<br />

M. A continuous curve in M connecting two points x, y of M <strong>and</strong> of length ϱ(x, y)<br />

is called a geodesic segment with endpoints x, y. It is non-trivial to show that, for<br />

sufficiently differentiable closed convex surfaces, or, more generally, for sufficiently<br />

smooth manifolds, this notion of geodesic segment coincides with the corresponding<br />

differential geometric notion. For a hint to a proof, see <strong>Gruber</strong> [423].<br />

Weyl’s Problem <strong>and</strong> Alex<strong>and</strong>rov’s Realization Theorem<br />

The following version of Weyl’s [1020] problem was studied in convex geometry,<br />

where a metric (d − 1)-sphere is a metric space homeomorphic to S d−1 .<br />

Problem 10.3. Let M be a metric (d − 1)-sphere. Is there a closed convex surface<br />

SinE d such that M <strong>and</strong> S are isometric, if both are endowed with their intrinsic<br />

metric? In other words, can M be realized by a closed convex surface?<br />

There exist many contributions to Weyl’s problem, mainly in differential geometry.<br />

In the following we state without proof Alex<strong>and</strong>rov’s [15] solution in the context of<br />

convex geometry.<br />

Let M be a metric 2-sphere endowed with its intrinsic metric. If two geodesic<br />

segments have a common endpoint it is possible to define a notion of lower angle<br />

between them. The intrinsic metric on M is then said to have positive curvature if<br />

each point of M has a neighbourhood with the following property: for each triangle<br />

in this neighbourhood with geodesic segments as edges, the sum of the lower angles<br />

between its edges is at least π. Alex<strong>and</strong>rov’s fundamental realization theorem now is<br />

as follows.<br />

Theorem 10.2. Let M be a metric 2-sphere with positive curvature. Then M is<br />

isometric to a closed convex surface S in E 3 where both M <strong>and</strong> S are endowed<br />

with their intrinsic metrics.<br />

Rigidity of <strong>Convex</strong> Surfaces <strong>and</strong> Pogorelov’s Rigidity Theorem<br />

The question arises, whether the convex surface S in Alex<strong>and</strong>rov’s realization theorem<br />

is unique up to rigid motions? In other words, if S <strong>and</strong> T are closed convex<br />

surfaces in E 3 which are isometric if both are endowed with their intrinsic metric, do<br />

they coincide up to rigid motions? If this is the case, S is called rigid.<br />

This problem for polytopal surfaces goes back to antiquity, see the discussion<br />

in Sect. 17.1. In the context of differential geometry the first pertinent result seems

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!