Applications of the Balanced Method to Stochastic Differential Equations in Filtering

The paper studies the application of the balanced method in hidden Markov chain filtering, an important practical area that requires the strong numerical solution of stochstic differential equations with multiplicative noise. Numerical experiments are conducted to enable comparisons between the balanced method and standard alternative methods in the context of filtering. Both the mean global error and the sample path properties of the approximate solutions are compared in a numerical study.