Let K⊂ Rⁿq, T⊂ Rⁿₚ be two bounded strictly convex bodies (open subsets) with C⁶-smooth boundaries. We consider the product K× T⊂ R²ⁿp,q equipped with the standard symplectic form ω=∑ⱼ₌₁ⁿdqⱼ dpⱼ. The $(K,T)$-billiard orbits are continuous curves in the boundary ∂(K× T) whose intersections with the open dense subset (K×∂ T)∪(∂ K× T) are tangent to the characteristic line field given by kernels of the restrictions of the symplectic form ω to the tangent spaces to the boundary. For every (q,p)∈ K× ∂ T the characteristic line in T(q,p) R²ⁿ is directed by the vector ( n(p),0), where n(p) is the exterior normal to Tₚ∂ T, and similar statement holds for (q,p)∈∂ K× T. The projection of each $(K,T)$-billiard orbit to K is an orbit of the so-called T-billiard in K. In the case, when T is centrally-symmetric, this is the billiard in Rⁿq equipped with Minkowski Finsler structure "dual to T", with Finsler reflection law introduced in a joint paper by S.Tabachnikov and E.Gutkin in 2002. Studying $(K,T)$-billiard orbits is closely related to C.Viterbo's Symplectic Isoperimetric Conjecture and the famous Mahler Conjecture in convex geometry. We study the special case, when the T-billiard reflection law is the projective law introduced by S.Tabachnikov, i.e., given by projective involutions of the projectivized tangent spaces Tq Rⁿ, q∈∂ K. We show that this happens, if and only if T is an ellipsoid, or equivalently, if all the T-billiards are simultaneously affine equivalent to Euclidean billiards. As an application, we deduce analogous results for Finsler billiards.
No takes yet. Share an insight, caveat, or question.
Alexey Glutsyuk (2024) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: