Nonautonomous chain rules in BV with Lipschitz dependence (Q506247)

The author recalls first some recent theorems on the nonautonomous chain rule formula in BV and proves that, in the scalar case, the nonautonomous chain rule formula for the distributed derivative of the composite function \(v(x)=B(x,u(x))\), with \(B(.,t)\) and \(u:\mathbb{R}^n\to \mathbb{R}\) of bounded variation, holds also by assuming a Lipschitz continuity of \(B(x,.)\), instead of \(C^1\) dependence. She obtains a nonautonomous chain rule formula in BV for the distributional derivative of the composite function \(v(x)=B(x,u(x))\), where \(B(.,t)\) has bounded variation and \(B(x,.)\) is Lipschitz continuous and differentiable in \(\mathbb{R}\backslash \mathcal{M}_0\) (\(\mathcal{M}_0\) independent of \(x\) with \(\mathcal{L}^1(\mathcal{M}_0)=0\)). The author requires the existence of a Radon measure \(\sigma\) bounding from above all measures \(|D_xB(.,t)|\), uniformly with respect to \(t\in \mathbb{R}\). For any \(u\in BV_{\mathrm{loc}}\) the composite function \(v(x) =B(x, u(x)\) belongs to \(BV_{\mathrm{loc}}\) and there exists of a countably \(\mathcal{H}^{N-1}\)-rectifiable set \(\mathcal{N}\), independent of \(t\) and containing the jump set of \(B(.,t)\) for every \(t\in \mathbb{R}\), such that the jump set of \(v\) is contained in \(\mathcal{N}\cup J_u\). Moreover, one obtains the following rule for the distributional derivative of \(v\): \(Dv(x) = D_x B(x,t)|_{t=u(x)}+(\partial_tB)(x,u).\tilde Du+ [B ^*(x,u^+)-B^*(x,u^-)]\nu_u\mathcal{H}^{N-1} \lfloor \;\mathcal{N}\cup J_u\) in the sense of measure, where the measure \(D_xB(x,t)\), depending on the parameter \(t\), is computed in \(t=u(x)\) in a suitable sense. The last section of the paper contains the proof of this result.