Comlab

When studying invariant quantities and stability of discretization schemes for time-dependent differential equations(ODEs), Backward error analysis (BEA) has proven itself an invaluable tool. Although the established results give very accurate estimates, the known results are generally given for "worst case" scenarios. By taking into account the structure of the differential equations themselves further improvements on the estimates can be established, and sharper estimates on invariant quantities and stability can be established. In the talk I will give an overview of BEA, and its applications as it stands emphasizing the shortcoming in the estimates. An alternative strategy is then proposed overcoming these shortcomings, resulting in a tool which when used in connection with results from dynamical systems theory gives a very good insight into the dynamics of discretized differential equations.

We consider the problem of computing ratings using the results of games (such as chess) played between a set of n players, and show how this problem can be reduced to computing the positive eigenvectors corresponding to the dominant eigenvalues of certain n by n matrices. There is a close connection with the stationary probability distributions of certain Markov chains. In practice, if n is large, then the matrices involved will be sparse, and the power method may be used to solve the eigenvalue problems efficiently.