A note on Euler's approximations (Q1266320): Difference between revisions

The author considers the stochastic differential equation in a domain \(D\) of \(R^d\): \[ dx(t)= b\bigl(t,x(t) \bigr)dt+ \sigma\bigl(t,x(t) \bigr) dw(t), \quad x(0)= \xi,\tag{1} \] and defines Euler's polygonal approximations as the process \(x_n(t)\) satisfying \[ dx_n(t)= b\biggl(t,x_n \bigl(\kappa_n (t) \bigr) \biggr)dt +\sigma \biggl(t,x_n \bigl(\kappa_n(t) \bigr)\biggr) dw(t), \quad x_n(0)=\xi,\tag{2} \] where \(\kappa_n(t):=t^n_i:=i/n\) for \(t\in [t^n_i, t^n_{i+1})\). It is shown that if the drift satisfies the monotonicity condition and the diffusion coefficient is locally Lipschitz, then \(x_n(t)\) converges to a process \(x(t)\), almost surely, uniformly in \(t\) in bounded intervals and the process \(x(t)\) is the unique solution of equation (1).