Fredholm and regularity theory of Douglis-Nirenberg elliptic systems on \({\mathbb{R}}^{N}\) (Q663258)

Denote by \(\mathcal P:=(p_{kl})\), \(1\leq k,\ell\leq r\), an \(r\times r\) Douglis-Nirenberg elliptic system of differential operators on \(\mathbb R^N\), so that either \(p_{k\ell}=0\) or \(p_{k\ell}:=\sum_{|\alpha|\leq d_{k\ell}}a_{k\ell\alpha}(x)\partial^{\alpha}\), where \(d_{k\ell}\in\mathbb N\cup\{0\}\) and the coefficients \(a_{k\ell\alpha}\) are complex-valued, with D-N numbers \(\mathbf s:=(s_1,\dots,s_r)\in\mathbb Z^r\) and \(\mathbf t:=(t_1,\dots,t_r)\in\mathbb Z^r\), i.e., \(d_{k\ell}\leq s_k+t_{\ell}\). If \(p_{k\ell}=0\), then \(d_{k\ell}:=-\infty\). Under suitable assumptions on the coefficients, \(\mathcal P\) is a well defined operator from \(W^{\mathbf t,p}:=\prod_{i=1}^rW^{t_i,p}\) into \(W^{-\mathbf s,p}:=\prod_{i=1}^rW^{-s_i,p}\) for every \(p\in (1,\infty)\), where \(W^{m,p}:=W^{m,p}(\mathbb R^N)\) is the classical Sobolev space of order \(m\) on \(L^p:=L^p(\mathbb R^N)\). In fact, the assumptions ensure that \(p_{k\ell}\) maps \(W^{t_{\ell},p}\) continuously into \(W^{-s_k,p}\) for every \(p\in (1,\infty)\). Hence \(\mathcal P\in\mathcal L(W^{\mathbf t,p},W^{-\mathbf s,p})\), where \(\mathcal L(X,Y)\) denotes the space of all bounded linear operators from the Banach space \(X\) into the Banach space \(Y\). The author introduces the matrix symbol \[ \sigma_{\mathcal P}(x,\xi):=\Big(\sum_{|\alpha|\leq d_{k\ell}}a_{k\ell\alpha}(x)i^{|\alpha|}\xi^{\alpha}\Big)_{k,\ell=1}^r \] and sets \(M:=\sum^r_{k=1}(s_k+t_k)\). Then, the main Fredholmness theorem reads as follows. {Theorem.} Let \(p\in (1,\infty)\) be given. The operator \(\mathcal P\in\mathcal L(W^{\mathbf t,p},W^{-\mathbf s,p})\) is Fredholm if and only if there are constants \(c>0\) and \(R\geq 0\) such that \[ |\det \sigma_{\mathcal P}(x,\xi)|\geq c(1+|\xi|^2)^{\frac M2}, \] for every \((x,\xi)\in\mathbb R^N\times\mathbb R^N\) such that \(|x|^2+|\xi|^2\geq R^2\). After the theorem, various invariance properties are established (index, null space, etc.), with respect to \(p\) and the D-N numbers. In turn, with some work, invariance leads to a regularity theory more general than what can be obtained by the method of differential quotients.