Weak and strong approximations of reflected diffusions via penalization methods (Q1940231): Difference between revisions

Reflected diffusions in a convex domain \(D\) with measurable coefficients are approximated by diffusions that penalize excursions outside \(D\), with a penalty proportional to the deviation from \(D\). In the infinite penalty limit, the latter are shown to approximate the weak, resp., strong, solutions of the former in law, resp., in probability, under conditions which are very general. Under stronger conditions, rate of convergence results are obtained in terms of bounds on the \(p\)-th moment, \(p \geq 1\), of the maximum distance between the trajectories on a compact time interval. Better bounds are shown to hold when \(D\) is a polyhedron.