Complexity and effective dimension of discrete Lévy areas (Q883328): Difference between revisions

To simulate stochastic differential equations, there exist several Euler- or Runge-Kutta-like methods which are analogues of well-known approximation schemes in the nonstochastic case. In the multidimensional case, there appear mixed terms corresponding to special random variables called Lévy stochastic area terms. In the present paper three approximation methods for such random variables are compared with respect to computational complexity and effective dimension. Each method is related to a different decomposition of the covariance matrix of finite-dimensional projections of the linear Brownian path.