This study presents an approach for distinguishing a driver from a response system. The proposed method can be applied to both unidirectional or bidirectional interactions, and to identical or structurally different systems. Compared with most previously proposed schemes, the present method is so "simple" that the driver-response relationships can generally be detected using a direct graphic way. On the other hand, quantitative estimation is also developed using the idea of the correlation

more »

... sion. A major issue in nonlinear dynamics and nonlinear time series analysis is the interplay among complex dynamic systems. Of particular interest is how to assess the interaction between two systems via analyzing the interrelation between their output signals. Up to now, many previously proposed schemes have either only been applicable to the generalized synchronization regime, or have had difficulty assessing the coupling direction in structurally different systems. In particular, all suggested schemes are so "complicated" that even for the simplest case, i.e., the case involving unidirectionally coupled identical systems, large-scale computations are required to distinguish driver-response relationships. This study suggests a method of addressing this problem. The proposed method can be applied to both weak and strong interactions, and to both identical and structurally different systems. Furthermore, the proposed method is sufficiently "simple" that, in most cases, it is even possible to directly estimate the driver-response relationships graphically. Additionally, a systematic technique based on the correlation dimension is devised to quantify the coupling asymmetry.