Semilinear \(p\)-evolution equations in Sobolev spaces (Q264480): Difference between revisions

The authors are interested in \(H^\infty\) well-posedness of the Cauchy problem in \([0,T]\times\mathbb{R}\) for \(p\)-evolution equations \[ P(t,x,D_t,D_x) u= D^m_t u+\sum^m_{j=1} a_{jp,j}(t) D^{jp}_x D^{m-j}_t u+ \sum^m_{j=1} \sum_{\alpha< jp} a_{\alpha,j}(t, x) D^\alpha_x D^{m-j}_t u= f(t,x). \] So, in the principal part in the sense of Petrovsky we have time-dependent coefficients. The Lax-Mizohata theorem tells us that the roots of the principal part in the sense of Petrovsky should be real. But coefficients depending on space variables, too, are allowed in the other terms. One already knows from Schrödinger operators that in convection terms one has to assume some decay behavior with respect to \(x\) from the influence of coefficients on phase functions. The \(p\)-evolution operators are related to Schrödinger operators, so there is a requirement that not only coefficients of ``terms of lower order'' should decay in \(x\) in a suitable way, but also the derivatives up to a certain order. The main reason is that in pseudo-calculus derivatives in \(x\) have some influence. The authors propose a hierarchy of decay assumptions for coefficients. In a first step they prove an \(H^\infty\) well-posedness result for special diagonal systems with remainders belonging to a suitable symbol class. The main step of the proof for the starting Cauchy problem is to develop a factorization and diagonalization procedure which reduces the given Cauchy problem to a system with the required properties. The authors propose a multi-step factorization procedure. Finally, they introduce a suitable energy and propose a multi-step diagonalization procedure. After producing remainders with suitable properties, an a priori estimate gives the \(H^\infty\) well-posedness. It would be interesting to understand the necessity of decay conditions for the coefficients in the future.