Moment estimation in uncertain differential equations based on the Milstein scheme

Abstract

Difference schemes are needed for approximating uncertain differential equations in applications. This paper mainly derives a new difference scheme called the Milstein scheme. It is theoretically shown that the Milstein scheme is superior to the previous Euler scheme. Then the Milstein scheme is applied to the method of moments so that the estimated uncertain differential equation fits the observations better. Moreover, the bias function is introduced to assess the precision of the estimation method. Finally, some numerical examples are given to verify the performance of both schemes and minimum cover estimation.