Algorithms for symmetric differential systems (Q5956408): Difference between revisions

A fundamental operation with overdetermined systems of partial differential equations is their completion, i.e., the addition of all integrability conditions. A number of different approaches to this problem exists; here a differential algebraic version with the Kolchin-Ritt algorithm as basis is employed. This is a simple critical pair completion algorithm analogous to the Buchberger algorithm for determining Gröbner bases. The required computations tend to be rather tedious and sometimes blow up, so that they cannot be completed even with a computer algebra system. A special situation appears, if the system to be completed is invariant under the action of a symmetry group. The algebraic completion requires the choice of a ranking for the derivatives which typically breaks the symmetry and increases the expression swell. In order to avoid this problem, the author rewrites the system in terms of differential invariants using the recently developed moving frame method of \textit{M. Fels} and \textit{P. J. Olver} [Acta Appl. Math. 51, 165-213 (1998; Zbl 0937.53012) and 55, 127-208 (1999; Zbl 0937.53013)]. She formulates a variant of the Kolchin-Ritt algorithm applicable in this situation and uses it for the analysis of the \((2+1)\)-dimensional d'Alembert-Hamilton system. This adaption of the Kolchin-Ritt algorithm is nontrivial, as the invariant differential operators associated with the moving frames do not commute in general. It is described for systems with at most polynomial nonlinearities. Furthermore, it is assumed that the action of the Lie algebra is rational and the moving frame polynomial in order to be able to perform effectively all computations. Finally, a brief description of a Maple implementation and a comparison with earlier work by \textit{I. G. Lisle} [Equivalence Transformations for Classes of Differential Equations, Ph.D. thesis, University of British Columbia (1992)] is given. The given examples demonstrate the high potential of the approach; the involved expressions become significantly smaller and are much easier to handle. A caveat is that it relies on several plausible but yet unproven conjectures of the author.