Explicit solutions to Poncelet's porism (Q417490)

Poncelet's porism states that if for two conics there is an \(N\)-gon, \(N\geq 3\), such that it is inscribed in one of them and circumscribed about the other then there exists an infinite number of such \(N\)-gons and one vertex of an \(N\)-gon may be located on the first conic arbitrarily. The first solutions to Poncelet's porism were proposed by W. Chapple, before Poncelet's theorem became known, who formulated an equation which linked the radii of the circles with the distance between their centres guaranteeing the result for triangles. Later, N. Fuss derived equations for \(N=3,4,5,6,8\). Many other results of the mathematicians of the 19-th and the first half of the 20-th century who solved the porism by algebraic equations for different values of \(N\) can be found in the web site of E.W. Weinstein. In this paper the case of two circles is considered. The author gives a recurrent procedure based on a biquadratic equation in order to determine if there exists a closed \(N\)-sided polygon interscribed between them. The number of arithmetic operations is near \(\log_2N\). The procedure may also be applied to determine whether a Hankel determinant of order \(N/2\) having a special form is equal to zero. For circles with rational coefficients, it is proved that the possible numbers \(N\) which are sides of interscribed polygons form the set \(\{3,4,5,6,8,10,12\}\), which is a subset of the Mazur's set of orders of torsion points of rational elliptic curves. The author also presents possible generalizations for ellipses and hyperbolas.