Non-Euclidean geometries based on ordinary high-order differential system (Q1358357)

According to the author, if one looks at the development of non-Euclidean geometry, starting with the work of Nicolai Lobachevskij, the first natural expansion of non-Euclidean geometries (taken as the class of Riemannian spaces), is the class of affinely connected spaces that preserve all the attributes of Riemannian spaces, based on parallel displacement, but loose the metric itself. A common feature of all these types of geometries is that they are based on an affine connection. The coefficients of this connection are the Christoffel symbols, which are the coefficients of a system of second order ordinary differential equations that determines self-parallel geodesics. Therefore it is possible to interpret all these geometries as geometries of similar systems of second order ordinary differential equations and, simultaneously, to complete in a natural way the process of extension of the class of non-Euclidean geometries within the framework of fiber bundles of first order frames. An ordinary second order differential system on a smooth manifold \(V_{n}\) is an arbitrary section \(f:T(V_{n})\rightarrow T^{2}(V_{n})\) under the condition that the coordinate representation of the system will admit a reduction. We can then project the above section into the section \(\varphi :S(V_{n})\rightarrow S^{2}(V_{n})\) of the fiber bundle of elements of second order contact over the bundle \(S(V_{n})\) of linear elements, which is an ordinary second order differential system with \(n-1\) equations. A reducible system \(f\) has the same family of integral curves as its reduced subsystem \(\varphi\), and the only additional equation in the system \(f\) only determines a two-parameter family of parametrizations on the integral curves. In the present article a detailed proof is given for the fact that on a manifold \(V_{n}\) systems of ordinary differential equations of order three and higher, which are reducible, generate geometries on the manifold that are a direct high order analogue of the first order geometrical structures corresponding to systems of second order equations. The integral curves of equations of order higher than the second play the role of generalized geodesics. The proof of this theorem presents some technical difficulties since the search for a geometry invariantly related to a given differential system is already nontrivial for third order systems, and becomes very complicated for higher order systems. In order to overcome these problems, the author uses the Cartan-Laptev method.