Skip to Main Content
Decentralized and self-organizing mechanisms present promising approaches to deal with the massive amount of data generated by different components in the smart grid. In this work we present a self-organized approach that is able to take decisions locally and in a distributed way. Our approach distinguishes between an infrastructure level and a decision level. The middleware processes running on the infrastructure level, which provides services such as routing, data filtering and aggregation. The decision level defines an explicitly and declaratively represented dynamic meta model that provides the semantics for the infrastructure level processes, which support interoperability. Additionally, this level runs processes that design, supervise and control agents on the infrastructure level. The levels coupled together through feedback loop that ensures that relevant changes on the infrastructure level are reflected in the decision level and vice versa. The corresponding reflection principles provide the basis for the implementation of the self-organizing mechanisms that govern the overall system. We distinguish two complementary categories of quality metrics for evaluating self-organizing systems. The first category explores the performance of the algorithm with respect to the target application such as delay, path length and success rate. In the context of self-organizing systems, it is important to evaluate additional kind of metrics that do not depend necessarily on the application rather than on self-organizing characteristics of the algorithm. In this paper we focus on the second type, and particularly on self-organizing properties of the algorithms.