On a condition for the union of spherical caps to be connected (Q1869762): Difference between revisions

Let \(\mathcal C= \{C_1, C_2,\dots, C_n\}\) be a family of spherical caps on the \(d\)-dimensional unit sphere \(S^d\), not necessarily of the same size. It is assumed that their centers are all distinct. Let \(v_i\) denote the center of \(C_i\). The intersection graph \(\Omega(\mathcal C)\) of \(\mathcal C\) is the graph whose vertices are \(v_1,v_2,\dots,v_n\) and two vertices \(v_i,v_j\) are adjacent (\(v_i\) is said to be a neighbor of \(v_j\)) only when \(C_i\cap C_j=\varnothing\). In the family \(\mathcal C\), a cap \(C_i\) is called an extremal cap if there is a hyperplane \(H\) passing through the origin and \(v_i\) such that all the neighbors of \(v_i\) in \(\Omega(\mathcal C)\) lie in the same side of \(H\), allowing that some of them lie on \(H\). The author proves that in the two-dimensional case \(d = 2\), if \(\mathcal C\) contains no extremal cap, then \(\Omega(\mathcal C)\) is connected, and if furthermore every cap in \(\mathcal C\) is smaller than a hemisphere, then \(\Omega(\mathcal C)\) is 2-connected. Surprisingly, the same assertion is no longer true for \(d\geqslant 3\).