Convergence of discretized stochastic (interest rate) processes with stochastic drift term. (English) Zbl 0915.60064

The authors study the stochastic differential equation \[ dX_{s}= (2\beta X_{s}+\delta_{s})ds+ g(X_{s})dB_{s}, \quad s\in \mathbb{R}^{+}, \] with \(X_0 \geq 0\), \(\beta \leq 0\), where \(g\) is a function vanishing at zero which satisfies the Hölder condition: \(| g(x)-g(y)|\leq b| x-y|^{1/2}\) and \(\delta_{s}\) is a measurable and adapted stochastic process such that \(\int_0^{t}\delta_{u} du<\infty\) a.e. The authors discuss the Euler discretization scheme for this stochastic differential equation with a drift term which may depend on a stochastic process with random correlation. It is shown that the approximating solution converges in \(L^1\)-supnorm and \(H^1\)-norm to the solution of this differential equation. The authors find conditions under which the Euler scheme strongly converges with order \(v=0.5\) at time \(T\). The authors note that this stochastic differential equation may be applied in finance.


60H10 Stochastic ordinary differential equations (aspects of stochastic analysis)
62P20 Applications of statistics to economics
Full Text: DOI