A precise definition of reduction of partial differential equations (Q1307258)

By reduction in the title here is meant Lie symmetry reduction of partial differential equations to ordinary differential equations, more specifically reduction under conditional symmetries. Such types of symmetry came into practical use in the late eighties through the work of \textit{P. J. Olver} and \textit{P. Rosenau} [Phys. Lett. A114, 107-112 (1986)] and \textit{V. I. Fushchich} and \textit{I. M. Tsifra }[J. Phys. A 20, L45--L48 (1987; Zbl 0663.35045)] based on an earlier idea of \textit{G. W. Bluman} and \textit{J. D. Cole} [J. Math. Mech. 18, 1025-1042 (1969; Zbl 0187.03502)]. A partial differential equation, \(L=0\) is conditionally invariant under a set of involutive operators \(\{Q_a|a=1\ldots n\}\) if the application of their prologations to \(L\) vanishes not on the submanifold of jet space defined by the equation but only on its intersection with those associated with the \(Q_a\)'s. The authors point out that in practice the idea is more easily used for equations with only two independent variables, and their aim is to provide a basis for a less ad hoc treatment of higher dimensional equations. They achieve this by constructing an ansatz adapted to the general solution of the symmetry defining system: \(Q_a\equiv 0\). This is a clear, straightforward treatment in which the reduction ansatz implies the presence of a conditional symmetry and vice versa. They apply their results to systems if the form \(\square u=F(u)\) where \(\square\) is the d'Alembertian in four dimensions, which reduce to d'Alembert-Hamilton systems. They consider special cases and recover some new solutions.