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

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

17 Rigidity 297<br />

in E k+1 with small spheres centred at the vertices <strong>and</strong> applying the result in S k ,<br />

the polytopal surface turns out to be rigid at each vertex. This immediately yields<br />

Cauchy’s theorem in E k+1 . Since the rigidity theorem holds in E 3 it thus holds in<br />

E d for all d ≥ 3. See Alex<strong>and</strong>rov [16] <strong>and</strong> Pogorelov [804]. For more information<br />

compare Connelly [218].<br />

Flexible Non-convex Polytopal Spheres <strong>and</strong> the Bellows Conjecture<br />

Cauchy’s rigidity theorem implies the following. A closed convex polytopal surface<br />

in E 3 with rigid facets <strong>and</strong> hinges along the edges cannot be flexed. Also the nonconvex<br />

example of Legendre of Fig. 17.1 does not admit a flexing. More generally,<br />

Euler [311] conjectured that<br />

a closed spatial figure allows no changes as long as it is not ripped apart.<br />

Bricard [167] gave an example of a flexible octahedral surface in E 3 . Unfortunately,<br />

it suffers from the defect that the surface is not embedded, that is, it has<br />

self-intersections. The next step of this story is due to Gluck [380] who showed<br />

that the closed, simply connected polytopal embedded surfaces in E 3 are generically<br />

rigid, that is, at least the large majority is rigid. Finally, Connelly [217], surprisingly,<br />

specified a flexible, embedded polytopal sphere in E 3 . A simpler example is due to<br />

Steffen [954].<br />

The so-called bellows conjecture asserts that the volume of a flexible polytopal<br />

sphere does not change while flexing. Sabitov’s [871] affirmative answer is based on<br />

an interesting formula in which the volume is expressed as a polynomial in terms of<br />

edge-lengths. This formula may be considered as a far-reaching extension of formulae<br />

of Heron for the area of a triangle <strong>and</strong> Euler for the volume of a tetrahedron. For<br />

a survey see Schlenker [889].<br />

17.2 Rigidity of Frameworks<br />

A framework is a system of rods in E d with joints at common endpoints such that<br />

the rods can rotate freely. A basic question is to decide whether a given framework<br />

is rigid or flexible.<br />

Early results on frameworks in the nineteenth century are due to Maxwell [700],<br />

Peaucellier [787], Kempe [573], Bricard [167]. Throughout the twentieth century<br />

<strong>and</strong>, in particular, in the last quarter of it, a multitude of results on frameworks were<br />

given. Amongst others, these results deal with rigidity <strong>and</strong> infinitesimal rigidity, with<br />

stresses <strong>and</strong> self-stresses.<br />

In this section a result of Asimow <strong>and</strong> Roth [41] will be presented, showing that<br />

a framework consisting of the edges <strong>and</strong> vertices of a proper convex polytope in E 3<br />

is rigid if <strong>and</strong> only if all facets are triangular. We follow Roth [859].<br />

Definitions<br />

An abstract framework F consists of two sets, the set of vertices V = V(F) =<br />

{1,...,v} <strong>and</strong> the set of edges E = E(F), the latter consisting of two-element subsets

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

Saved successfully!

Ooh no, something went wrong!