Given an open, bounded and connected set $\mathrm{\Omega }\subset {\mathbb{R}}^n$ with Lipschitz boundary and volume $\left|\mathrm{\Omega }\right|$, we prove that the sequence ${\mathcal{F}}_k$ of Dirichlet functionals defined on $H^1\left(\mathrm{\Omega };{\mathbb{R}}^d\right)$, with volume constraints $v^k$ on $m\ge 2$ fixed level-sets, and such that ${\sum }_{i=1}^m{v}_i^k<\left|\mathrm{\Omega }\right|$ for all $k$, $\mathrm{\Gamma }$-converges, as $v^k\to v$ with ${\sum }_{i=1}^m{v}_i^k=\left|\mathrm{\Omega }\right|$, to the squared total variation on $BV\left(V;{\mathbb{R}}^d\right)$, with $v$ as volume constraint on the same level-sets.

Questo articolo considera una successione di equazioni differenziali a derivate parziali non lineari in forma di divergenza del tipo $$-\text{div}\left(Q^{ϵ}G\left(x,N^{ϵ}\nabla u\right)\right)=f^{ϵ},$$ in un dominio limitato $\mathrm{\Omega }$ dello spazio $n$-dimensionale; $Q^{ϵ}=Q^{ϵ}\left(x\right)$ e $N^{ϵ}=N^{ϵ}\left(x\right)$ sono matrici con coefficenti limitati, $N^{ϵ}$ e è invertibile e la sua matrice inversa $R^{ϵ}$ ha anche coefficenti limitati. La non linearità è dovuta alla funzione $G=G\left(x,\xi \right)$; la condizione di crescita, la monotonicità e le ipotesi di coercitività sono modellate sul $p$-Laplaciano, $1<p<\mathrm{\infty }$, ed assicurano l'esistenza di...

In this paper we consider two-dimensional quasilinear equations of the form $\text{div}\left(a\left(\left|\nabla u\right|\right)\nabla u\right)+f\left(u\right)=0$ and study the properties of the solutions u with bounded and non-vanishing gradient. Under a weak assumption involving the growth of the argument of $\nabla u$(notice that $\text{arg}\left(\nabla u\right)$ is a well-defined real function since $\left|\nabla u\right|>0$ on ${\mathbb{R}}^2$) we prove that $u$ is one-dimensional, i.e., $u=u\left(\nu \cdot x\right)$ for some unit vector $\nu$. As a consequence of our result we obtain that any solution $u$ having one positive derivative is one-dimensional. This result provides...

We define a mapping which with each function $u\in L^{\mathrm{\infty }}\left(0,T\right)$ and an admissible value of $r>0$ associates the function $\xi$ with a prescribed initial condition ${\xi }^0$ which minimizes the total variation in the $r$-neighborhood of $u$ in each subinterval $\left[0,t\right]$ of $\left[0,T\right]$. We show that this mapping is non-expansive with respect to $u$, $r$ and ${\xi }^0$, and coincides with the so-called play operator if $u$ is a regulated function.