Morse index of a cyclic polygon (Q539205)

An \(n\)-linkage is a tuple of real positive numbers \(L=(\ell_1,\ldots, \ell_n)\) which can be realized as edge lengths of a planar polygonal chain (self-intersections allowed). A configuration \(P\) associated to a linkage \(L\) is a set of \(n\) ordered vertices \(p_i\) such that the length between consecutive vertices \(|p_i,p_{i+1}|\) is \(\ell_i\). The first two vertices are ``pinned down'' to factor out the action of \(SO_2\). The (moduli) space \({\mathcal M} (L)\) of all such configurations is a smooth closed manifold and its topology depends on the length of edges [\textit{M. Farber} and \textit{D. Schütz}, Geom. Dedicata 125, 75--92 (2007; Zbl 1132.52024)]. A configuration is called cyclic if all its vertices lie on a circle. It turns out -- by a beautiful result of the first author and \textit{G. Khimshiashvili} [J. Math. Sci., New York 158, No. 6, 899--903 (2009); translation from Zap. Nauchn. Semin. POMI 360, 238--245 (2008; Zbl 1193.52015)] -- that the signed area function \(A(P)\) of a configuration \(P\) is a Morse function on \({\mathcal M} (L)\) of which critical points are precisely the cyclic configurations. The paper at hand gives an explicit workable formula for the Morse index at such critical points. First of all coordinates are introduced so that the Hessian at a critical point \(P\) is computed. The sign \({\mathcal H }(P)\) of the determinant of this Hessian is then determined in terms of tangents of half the internal angles defined by the cyclic configuration and also in terms of the position of the center of the circle with respect to the edges, provided no such edge contains the center (Theorem 4.6). The final formula is given in Theorem 5.2, and it expresses the index at the critical point \(P\) in terms of the number of sign changes in the sequence \({\mathcal H }(P_3), {\mathcal H }(P_4),\cdots , {\mathcal H }(P_n)\), \(P=P_n\), with \(P_i\) the subconfiguration on vertices \((p_1,\ldots, p_i)\).