[<<][math][>>][..]Thu Jan 6 14:28:24 EST 2011

Called divided difference as it's defined as: d x[k] = x[k+1] - x[k] / T I.e. the differences are taken relative to the sampling period T. In most cases where T is fixed this can be normalized to T=1 by moving the scaling to the coefficient end, i.e. coefficients will become small when T approaches 0. However, in the exposition of [1] explicit mention of T is desired to be able to take the limiting case of T->0 to make the bridge to the continuous case, and the cases with different T. See also fixed step integration[2], Time Scale Calculus[3]. [1] md5://3e41deebd80a5b988de09b645c036e57 [2] http://en.wikipedia.org/wiki/Eulers_method [3] http://en.wikipedia.org/wiki/Time_scale_calculus

[Reply][About]

[<<][math][>>][..]