Some Mathematical And Numerical Questions Connected With First And Second Order Time Dependent Systems Of Partial Differential Equations

Abstract

There is a tendency to write the equations of general relativity as a first order symmetric system of time dependent partial differential equations. However, for numerical reasons, it might be advantageous to use a second order formulation like one obtained from the ADM equations. Unfortunately, the type of the ADM equations is not well understood and therefore we shall discuss, in the next section, the concept of wellposedness. We have to distinguish between weakly and strongly hyperbolic systems. Strongly hyperbolic systems are well behaved even if we add lower order terms. In contrast; for every weakly hyperbolic system we can find lower order terms which make the problem totally illposed. Thus, for weakly hyperbolic systems, there is only a restricted class of lower order perturbations which do not destroy the wellposedness. To identify that class can be very difficult, especially for nonlinear perturbations. In Section 3 we will show that the ADM equations, linearized around flat with constant lapse function and shift vector, are only weakly hyperbolic. However, we can use the trace of the metric as a lapse function to make the equations into a strongly second order hyperbolic system. Using simple models we shall, in section 4, demonstrate that approximations of second order equations have better accuracy properties than the corresponding approximations of first order equations. Also, we avoid spurious waves which travel against the characteristic direction. In the last section we discuss some difficulties connected with the preservation of constraints.