Mesoscopic Analysis of an Artiicial Chemistry with Self-organizing Topology

The analysis of artiicial life systems is often much harder than their construction, especially in population-based systems with a huge number of interacting individuals. Due to insuucient results of microscopic methods (e.g. monitoring the actions of single molecules) and macroscopic measurements (e.g. diversity, complexity) we developed a stepwise mesoscopic analysis method based on classiication and dynamic clustering. Knowledge about the system is accumulated in an iterative process, in which classiiers (feature detectors) extract information, which is used in turn to create new classiiers. The mesoscopic analysis is applied to an ar-tiicial chemistry with a self-organizing topology. The topo-logical structure is determined by the molecules themselves and is based on hashing which is a very eecient memory access method.