General order conditions for stochastic Runge-Kutta methods for both commuting and non-commuting stochastic ordinary differential equation systems (Q1294506)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | General order conditions for stochastic Runge-Kutta methods for both commuting and non-commuting stochastic ordinary differential equation systems |
scientific article |
Statements
General order conditions for stochastic Runge-Kutta methods for both commuting and non-commuting stochastic ordinary differential equation systems (English)
0 references
16 November 1999
0 references
This paper concerns Runge-Kutta methods for approximating the solution of the Stratonovich vector stochastic ordinary differential equation \[ dy= g_0(y)dt+ \sum^d_{j= 1} g_j(y)\circ dW_j, \] where the \(g_j\) are vector functions and the \(W_j\) are independent Wiener processes. It is shown that in the general case of this equation, if more than one Wiener process is present, the maximum strong order that can be attained by a stochastic Runge-Kutta method with random variables representing all order 1 and order 2 Stratonovich integrals is 0.5 irrespective of the number of stages in the method. However, it is also shown that under certain conditions order 1 can be attained if the commutative property \(g_i'(y) g_j(y)- g_j'(y) g_i(y)= 0\) is studied for all \(i\), \(j\), \(y\). Numerical results for three known methods using different stepsizes illustrate the 0.5 order restriction.
0 references
order conditions
0 references
numerical examples
0 references
Stratonovich vector stochastic ordinary differential equation
0 references
Wiener processes
0 references
stochastic Runge-Kutta method
0 references
0 references
0 references
0 references