Given a convex parallel opposite sides polygon (CPOSP) P1P2…P2n we can define:
Figure 1. AE (red) and CSS (green) of a CPOSP
Note that:
In this Geogebra applet, the vertices of the AE and of the CSS are marked red whenever they turn into cusps (as you play with the outer polygon).
This is a Java Applet created using GeoGebra from www.geogebra.org – it looks like you don’t have Java installed, please go to www.java.com
For non-degenerate positions (see [1] for proofs):
If we define λi+½ (i=1,2,…,2n) by the equations
Di+½ = Pi + λi+½·(Pi+n – Pi) = Pi+1 + λi+½·(Pi+1+n – Pi+1)
then we have
Di+½ – Di-1+½ = (λi+½ – λi-1+½)(Pi+n – Pi)
Now, in a CPOSP, the great diagonals Pi+n – Pi “turn” in the same orientation as the polygon itself. Therefore the triangles Di-1+½ Di+½ Di+1+½ and Pi-1PiPi+1 have opposite orientations if and only if
(λi+½ – λi-1+½)(λi+1+½ – λi+½)<0
In short
Di+½ is a cusp of the CSS if, and only if, λi+½ is a local extremum of the cyclic sequence {λi+½}
[1] M. Craizer, R. Teixeira and M. Horta, “Parallel Opposite Side Polygons”, preprint.