Some elementary questions in the calculus of variations (Q2039368): Difference between revisions

The author considers a general functional \(\mathcal{F}(y)= \int_{a}^{b}f(t,y(t),y^{\prime }(t))dt\) with a Lagrangian function \( f:[a,b]\times \mathbb{R}\times \mathbb{R}\rightarrow \mathbb{R}\) and the minimization of \(\mathcal{F}\) on a set of functions belonging to different spaces: \(AC([a,b])\), \(W^{1,1}(a,b)\), Lip\(([a,b])\) or \(C^{1}([a,b])\) and satisfying \(y(a)=y_{a}\) and \(y(b)=y_{b}\). The author recalls the Lavrentiev phenomenon which may occur \[\inf_{\substack{y\in AC([a,b]),\\ y(a)=y_{a},\\ y(b)=y_{b}}}\mathcal{F}(y)<\inf_{\substack{y\in \mathrm{Lip}([a,b]),\\y(a)=y_{a}, \\y(b)=y_{b}}}\mathcal{F}(y).\] The purpose of the paper is to exhibit some examples of the Lavrentiev phenomenon between \(AC/\mathrm{Lip}\) and \(\mathrm{Lip}/C^{1}\), the infima being really achieved. The author proves that the Lagrangians are slightly better than a Caratheodory function. The paper ends with uniqueness results of the minimizers under partial strict convexity hypotheses of the Lagrangian with respect to either the second or the third variable.