On sequential and parallel solution of initial value problems (Q911713)

It is well known that if the functionals are given in advance, independent of any particular f, then the information is called nonadaptive. On the other hand, if the functionals are chosen on the basis of all previously computed values then the informations are called adaptive. The main advantage of nonadaptive over adaptive information is that the former is well-suited for parallel computations, while the later is sequential. In this interesting paper the author shows that adaptation gives a much smaller error than nonadaptation. In fact, it is shown that the superiority of adaptation increases with the dimension. Several examples are given to illustrate the main ideas of the paper.