Nonlinear PDEs and existence questions

A local theorem constructs a solution on [0,T] with T possibly depending on the size and regularity of the initial data. A global theorem extends it to every finite positive time within the specified class. Regularity asks how many derivatives exist, in what sense, and where; uniqueness compares solutions in a stated class. Continuous dependence is an additional condition on the solution map. None of these terms is interchangeable. A classical solution may cease to exist while an entropy or energy weak solution continues, as in shock formation for conservation laws.

A continuation criterion identifies a quantity whose finite bound permits a local construction to restart. If that quantity becomes unbounded at a finite time, the theorem no longer guarantees classical continuation. The quantity might be a supremum norm, a derivative norm, or a scale-sensitive space-time norm, depending on the PDE. A growing plot does not prove blow-up, and bounded total energy does not automatically bound every derivative. In multiple spatial dimensions, concentration can make a derivative large while leaving an integral energy finite. The dimension and scaling of an estimate are therefore central to its strength.

A common existence strategy first regularizes the equation or projects it into finite-dimensional spaces. Solve those approximate problems and derive bounds independent of the approximation parameter. Compactness then extracts a subsequence converging in a topology strong enough to identify the limiting terms. Boundary conditions and initial data must also survive the limit. A priori means the estimate is derived for a candidate or approximate solution before knowing that a desired exact solution exists; it is not a claim that the solution is assumed into existence.

Nonlinear terms explain why the convergence topology matters. On a periodic interval, sin(nx) converges weakly to zero in L2, while its square converges weakly to one half rather than zero. Thus weak convergence of u_n alone does not imply that u_n² converges to u². Additional compactness, monotonicity, compensated structure, or an appropriate defect description may be needed. This is one reason a stable-looking numerical sequence and an energy bound do not themselves constitute a complete existence proof. The limiting equation is a mathematical conclusion that must be checked term by term.

Diffusion does not always prevent nonlinear growth. For u_t=κΔu+u² on a periodic domain, spatially constant positive data reduce the PDE to A'=A², which blows up at a finite time. The boundary choice matters: a nonzero constant is not compatible with homogeneous Dirichlet data, so that same example cannot be transplanted unchanged to a zero-temperature bounded wall. Other nonlinear equations have different mechanisms, including gradient blow-up, shocks, dispersion, and globally controlled solutions. No single example decides them all.

For three-dimensional incompressible Navier–Stokes, global finite-energy weak existence and global smooth regularity are distinct levels of theory; weak existence alone does not establish the smoothness demanded in the official Clay problem. The cited problem statement fixes data, domain alternatives, forcing conventions, and requested conclusions. Treat proposed solutions through that precise statement rather than a blanket claim that every fluid question is settled or unsolved. In the spectral lab, increase N at fixed p to inspect unresolved scales. The display has no nonlinear PDE update, compactness test, fluid dynamics, or blow-up detector; it only motivates why scale control needs an estimate.