Existence and regularity in the small in the calculus of variations (Q803527): Difference between revisions

The main result is the following: Let \((t_ 0,x_ 0)\) be any point in \({\mathbb{R}}\times {\mathbb{R}}^ n\) which admits a neighbourhood S having the following properties: the function (t,x,v)\(\to L(t,x,v)\) is Lipschitz on compact subsets of \(S\times {\mathbb{R}}^ n\) and for each (t,x) in S the function \(v\to L(t,x,v)\) is strictly convex on \({\mathbb{R}}^ n\). Then given any \(M>0\) there exist arbitrarily small positive \(\epsilon\) and \(\gamma\) such that for any a and b within \(\epsilon\) of \(t_ 0\) with \(a<b\), for any A and B within \(\epsilon\) of \(x_ 0\) with \(| A-B| \leq M(b- a)\), the variational problem of minimizing \(\int^{b}_{a}L(t,x,\dot x)dt\) over all absolutely continuous arcs x on [a,b] satsfying \(x(a)=A\), \(x(b)=B\), \(| x(t)-x_ 0| <\gamma\) (t\(\in [a,b])\) has at least one solution and all solutions are continuously differentiable on [a,b].