Reparametrizations and approximate values of integrals of the calculus of variations. (Q1412352)

The main result of the present paper is given by a reparametrization property of the following problem. Let be given the following functional \[ I(x)=\int_a^b L(x(t),x^\prime (t))\psi(t,x(t))\,dt, \] defined on functions \(x:[a,b]\to {\mathbb R}\) satisfying the conditions \(x(a)=A\) and \(x(b)=B\). In this setting, the authors prove that if \(x:[a,b]\to {\mathbb R}^N\) is an absolutely continuous function and \(C=\{ x(t): t\in [a,b]\}\), \(L:C\times {}{\mathbb R}^N\to {\mathbb R}\) is a continuous function with \(L(x,\cdot)\) convex, \(\psi :[a,b]\times C\to [c,+\infty)\) (\(c>0\)), then for every \(\varepsilon >0\), there exists \(x_\varepsilon\), a Lipschitz reparametrization of \(x\), such that \(x(a)=x_\varepsilon (a)\), \(x(b)=x_\varepsilon (b)\) and \[ I(x_\varepsilon)\leq I(x)+\varepsilon. \] The proof of the theorem is based on convex analysis and uses the notion of polar functions. As a consequence of the reparametrization property, the authors prove the absence of the Lavrentiev phenomenon under the assumption that the Lagrangian \(L\) is continuous, \(L(x,\cdot)\) is convex and \(\psi:[a,b]\to [c,+\infty)\) is continuous. Moreover, applications in some obstacle problem and in problems defined in several variables with radial invariance are also given.