GRADIS - Multiagent environment supporting distributed graph transformations

Abstract

Graph transformations are a powerful notation formally describing different aspects of modeled systems. Multiagent systems introduce distribution, parallelism and autonomous decision properties. In the paper a basic properties of the GRADIS agent's framework, joining of both approaches, are discussed. This framework supports splitting the graph, describing a problem, onto a few partial graphs, that can be maintained by different agents. Moreover, the multiagent's cooperation enables the application to the local graphs the graph transformation rules introduced for the centralizes graph; this permits us transfer all theoretical achievements of the centralized graph trans-formations to the distributed environment. The usefulness of the hierarchical graphs structure are and some examples of its usefulness are presented. © 2008 Springer-Verlag Berlin Heidelberg.