The functional definition of generalized geodesics (Q1577673)

Let \(M\) be a finite dimensional topological manifold, \(D\) an open subset of \(M^2\), containing the diagonal and \(I\) an open real interval and let \(f: D\times[0,1] \to M\) satisfy the equation \(f(a,b,(1-\gamma)\alpha+\gamma\beta)=f(f(a,b,\alpha),f(a,b,\beta),\gamma)\) and the boundary conditions \(f(a,b,0)=a\), \(f(a,b,1)=b\) for all \((a,b)\in D\), \((\alpha,\beta,\gamma)\in [0,1]^3.\) The author calls a function \(g: I\to M\) a generalized geodesic if, for somewhat restricted pairs \((\alpha, \beta) \in I^2\) and for all \(\gamma\in [0,1],\) the functional equation \(g((1-\gamma)\alpha+ \gamma\beta)=f(g(\alpha),g(\beta),\gamma)\). A proof is offered that this family of functions contains the geodesics in the usual sense but also other mappings.