Partitions of points into simplices with \(k\)-dimensional intersection. II: Proof of Reay's conjecture in dimensions 4 and 5 (Q5944013): Difference between revisions

This paper is a sequel to [part I of this paper, ibid., 733-743 (2001; Zbl 1007.52005)], reviewed above, in which the same author proved a generalization of Tverberg's theorem and used it to prove various lemmas (some of which are used in this paper) and proved Reay's conjecture for the special cases of partitions into 3 or 4 subsets. Here, the author (who has previously proved the conjecture in dimensions 2 and 3) extends it to dimensions 4 and 5. The concept of a \(k\)-lopsided oriented matroid on \(E\) is introduced, that is to say, one such that every subset of \(E\) of size less than or equal to \(k+1\) contains at most one positive circuit, and every positive circuit has at most \(k\) elements. Lemmas are proved giving necessary and sufficient conditions for a totally cyclic oriented matroid of rank \(d\) on an \(n\)-element set to be \(d\)-lopsided, for various pairs \((d,n)= (4,7),(4,8),(5,8)\), and \((5,9)\). The lopsidedness conjecture, introduced here, involves the structure of \(k\)-lopsided oriented matroids. The author proves it in certain cases; it is then used as part of the proof of Reay's conjecture when \(d=4\). While the lopsidedness conjecture would, if true, also provide the necessary mechanism to prove Reay's conjecture for \(d=5\), those cases have not been proved; an alternative method is used, considering several values of \(k\) and, for each, several configurations of points; the five-dimensional case, at length, follows.