Non-isotropic weighted \(L^{p}\) estimates for the \(\bar \partial\)-equation on piecewise smooth strictly pseudoconvex domains (Q1042812)

Let \(\Omega\) be a domain in \(\mathbb{C}^{n}\), verifying the following conditions: i) \(\Omega\) is the intersection of \(N\) strictly pseudoconvex domains \(\Omega_{j}=\left\{ \rho_{j}<0\right\}\) with \(C^{2}\) boundary \(b\Omega_{j}\); ii) (transversality condition): for every subsequence \(I\) of \(I=\left\{1,2,...,N\right\}\) the forms \(\partial_{\rho_{i_{1}}},\partial_{\rho_{i_{2}}},...,\partial_{\rho_{i_{k}}}\) with \(\left\{ i_{1},...,i_{n}\right\}=I\) are linearly independent on \(\bigcap_{\nu=1}^{l}U_{i_{\nu}}\), where the \(U_{i_{\nu}}\) are neighborhoods of \(b\Omega_{i_{\nu}}\). Let also \(\rho\) be defined by \(\dfrac{1}{\rho}=\mathop{\sum}\limits_{j\in I}\dfrac{1}{\rho_{j}}\). Then \(\rho\in C^{2}(\Omega)\), is comparable with the boundary distance, and \(\Omega=\mathop{\bigcup}\limits_{k\in I}\tilde{\Omega}_{k}\) where \(\tilde{\Omega}_{k}=\left\{ z\in\Omega|-\rho_{k}(z)=\mathop{inf}\limits_{i\in I}\left\{ -\rho_{i}(z)\right\} \right\}\). Finally, denote by \(\tilde{N}(z)=\left\{ k:z\in\tilde{\Omega}_{k}\right\}\). Then the main result of the AA is the following: Let \(\Omega\) be a bounded domain in \(\mathbb{C}^{n}\) verifying conditions (i)-(ii) (these domains are called piecewise strictly pseudoconvex). The equation \(\overline{\partial}u=f\), with \(f\) a \(\overline{\partial}\)-closed (0,1) form has a solution \(u\) in \(\Omega\) such that, for any \(\eta>0\), \(\alpha>0\), \(1\leq p<\infty\) it verifies the following estimates \[ \mathop{\int}\limits_{\Omega}|\rho(z)|^{\alpha-1}| u(z)|^{p}dV(z)\leq C_{p,\alpha,\eta}\mathop{\int}\limits_{\Omega}|\rho(z)|^{\alpha-1-\eta}\| f(z)\|_{\Omega}^{p}dV(z) \] and for any \(\eta>0\), \(\alpha>1\) \[ \mathop{\sup}\limits_{z\in\Omega}|\rho(z)|^{\alpha-1}| u(z)|\leq C_{\alpha,\eta}\mathop{\sup}\limits_{z\in\Omega}|\rho(z)|^{\alpha-1-\eta}\| f(z)\|_{\Omega} \] whenever the right-hand sides of the above inequalities are finite. Here the nonisotropic norm \(\|\cdot\|_{\Omega}\) of (0,1) forms is defined by \[ \| f(\zeta)\|_{\Omega}=|\rho(\zeta)\| f(\zeta)|+\mathop{\sup}\limits_{\nu\in\tilde{N}(\zeta)}|\rho_{\nu}(\zeta)|^{1/2}|\overline{\partial}\rho_{\nu}(\zeta)\wedge f(\zeta)|, \] \(\zeta\in\Omega\). The proof relies essentially on the Berndtsson-Andersson formula and, passing from the smooth case to the general case the transversality condition plays an essential role. The use of the weighted measure \(|\rho|^{\alpha-1}dV\) is technically delicate.