Abstract. In this article, we concern ourselves with a new concept for comparing the stability degree of two dynamical systems. By using the integral inequality method, we give a criterion which allows us to compare the growth rate of two Itô quasi-linear differential equations. It can be viewed as an extension of the Lyapunov criterion to the stochastic case.
IntroductionIt is known that studying the asymptotic behavior of dynamical systems is important both in theory and in application. Therefore, there are many works dealing with this topic and there exist a large amount of stability criteria for deterministic and stochastic systems. Among these criteria, the characteristic Lyapunov exponent is a powerful tool because it is important for explaining the chaos of the systems under consideration (see [1,2,10], etc.). We remark that studying the Lyapunov exponent of a function means comparing its growth rate with the growth rate of the exponential one. However, the class of exponential functions is rather simple and it does not contain much information on the behavior of the function considered. By requirement of technical problems, we have sometimes to replace this class by a larger one, say C, and compare this function with elements of C in order to know its behavior, especially at t = ∞. To realize this we introduce a concept of comparing the behavior of trajectories of solutions of two differential equations as follows: a system is said to be better than a given one in the class C in view of stability (we will say that this system is more stable than the given one) if, whenever all trajectories of the given system starting from a small neighborhood of the origin 0 belong to the corridor {(t, x): t ≥ 0, |x| ≤ q t } generated by a positive function q t ∈ C, then all trajectories of the second system starting from a suitable neighborhood of 0 must have the same property.Naturally, this raises a question: when is one system more stable than another? Of course, it is a difficult problem even in the deterministic case. In this article, we deal with a criterion for comparing two quasi-linear systems. This problem is encountered when we investigate a nonlinear stochastic system via its first linear approximation.The article is organized as follows. Section 2 gives a concept of comparing two differential stochastic equations. In Section 3, we are concerned with the regularity of a 2000 Mathematics Subject Classification. Primary 60H10; Secondary 34F05, 93E15.