P.L.-spheres, convex polytopes, and stress (Q1913692): Difference between revisions

A central problem in the theory of convex polytopes is the classification of the vectors \((f_0, \dots, f_{d - 1})\) which can occur as the numbers of faces of a \(d\)-polytope. The general problem still appears to be out of reach, but in the special case of simplicial polytopes, the characterization was proposed by the reviewer, and settled by \textit{L. Billera} and the author (sufficiency) [J. Comb. Theory, Ser. A 31, 237-255 (1981; Zbl 0479.52006)] and \textit{R. P. Stanley} (necessity) [Adv. Math. 35, 236-238 (1980; Zbl 0427.52006)]. The latter proof, though, used deep techniques from algebraic geometry, and it had to wait for the reviewer [Invent. Math. 113, No. 2, 419-444 (1993; Zbl 0803.52007)] to provide a proof (in the dual context, using weights on polytopes) which worked entirely within convexity. Another promising approach, though, involves (linear or affine) stress, which generalizes the classical stress of a bar-and-joint structure. This valuable article surveys the history of stress theory, introduces the recent developments due to the author which define stress on faces of greater dimension than one, and delineates the connexions with the (algebraically) dual theory of rigidity. Stress works in a more general context than do weights, and a still open question is whether the reviewer's characterization extends to simplicial spheres. The author describes much of how such extensions should behave; nevertheless, both Stanley's and the reviewer's proofs (loc. cit.) rely essential on convexity, and this still appears to be a formidable barrier. However, the author does extend earlier results on unbounded simple polyhedra, showing that the ``hyperplane section'' used in these proofs provides easier proofs of results on polytope pairs.