Comparison of dependency measures for the detection of mutual influences in organic computing systems

Abstract

Organic Computing (OC) postulates to tackle challenges arising from the increasing complexity in technical systems by means of self-organization and “life-like” properties. Recently, the arising complexity challenges have become even more severe due to the rapidly ongoing interweaving process–meaning that systems are directly and indirectly coupled and influencing each other. In order to be able to deal with such influences, they have to be identified and analyzed in the first place. Therefore, this paper reviews existing techniques to detect mutual influences among distributed entities at runtime. The goal is to compare the applicability of the varying concepts according to their applicability within Organic Computing systems. Therefore, we investigate two abstract (i.e., two simulated robots moving a crosscut saw and carrying a box) and a typical OC (i.e., smart camera network) example for evaluation purposes.