The A-Contractive Second-Order Multistep Formulas with Variable Steps
This paper investigates the A-contractivity in the maximum norm of second order multistep methods \[ \sum^{k}_{i=0}a_ ix_{n-i}=(t_ n- t_{n-1})\sum^{k}_{i=0}b_ i\dot x_{n-i} \] with arbitrarily variable steps. The principal result is the derivation of values \(b_ 0,b_ 1,...,b_ k\) to give a unique A-contractive second order method for any choice of non-positive \(a_ 1,a_ 2,...,a_ k\) with \(a_ 0=- \sum^{k}_{i=1}a_ i\). In this complete characterisation of second order A-contractive methods, the extreme cases in which \(a_ i=0\) (\(i\neq 0,i\neq j)\) are simply the trapezoidal rules \(x_ n-x_{n-j}=1/2(t_ n- t_{n-j})(\dot x_ n+\dot x_{n-j}).\) Results concerning contractivity apply not only to linear multistep methods but also to the corresponding one-leg methods and it is the development of stable and accurate algorithms of this type that motivates the work.
- Matrix valued versions of a result of von Neumann with an application to time discretization
- A class of contractive second derivatives multistep methods
- Multiplier and contractivity methods for linear multistep methods
- Refactorization of a variable step, unconditionally stable method of Dahlquist, Liniger and Nevanlinna
- scientific article; zbMATH DE number 179270 (Why is no real title available?)
- A-Contractivity of Linearly Implicit Multistep Methods
- Integrators for stiff systems with undamped oscillatory solutions
- Time integration algorithms with optimized controllable numerical dissipation for dynamic systems
- On a class of multistep methods with strong contractivity properties
This page was built for publication: The A-Contractive Second-Order Multistep Formulas with Variable Steps
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3327650)