System structure identification by analyzing elements behavior sequences with GRA-based ISM

Generally, it is a good starting-point and foundation for analyzing complex systems by establishing hierarchical structure, with interpretative structural modeling(ISM). It has been pointed out by many researchers that the traditional way to establish adjacency matrix in ISM has its limitation. In this paper the necessity and flexibility of integrating the grey relational analysis (GRA) for ISM is discussed. By incorporating GRA into ISM, an improved ISM, called GRA-based ISM is proposed to obtain system structure by analyzing the behavior sequences of its elements. The adjacency matrix is no longer generated by subjective assumption, but derived from the grey incidence matrix in GRA. With the pair-wise relations more objective and accurate, the system structure deduced from these relations is more rational. After the theoretical foundation of GRA-based ISM is analyzed, its analysis steps are explained in details, followed by an empirical study on Beijing socio-economy system structure to demonstrate its feasibility.