Persistence of attractors for one-step discretization of ordinary differential equations (Q2748868)

This paper deals with the behaviour of one step discretization methods with respect to the attractors of the original differential system. A basic result in this direction was proved by \textit{P. E. Kloeden} and \textit{J. Lorenz} [SIAM J. Numer. Anal 23, 986-995 (1986; Zbl 0613.65083)]. who showed that if the original differential equations have an attractor then the discretization possess absorbing sets which converge to the attractor as the time step tends to zero. However the existence of numerical attractors does not imply the existence of a nearby attractor for the original differential system. NEWLINENEWLINENEWLINEThe author studies this converse implication and proves that under some additional assumption it holds true. More precisely the upper limit of the numerical attractors is an attractor of the original differential equation if and only if the numerical one step scheme admits attracting sets with uniform attraction rate approximating this upper limit set. Furthermore if the numerical attractors themselves are attracting with uniform rate, then they converge to some set \(A\) if and only if this set \(A\) is an attractor of the original ordinary differential equation and an estimate of the rate of convergence may be derived.