A Degree-Based Decision-Centric Model for Complex Networked Systems

The system-level structure and performance of complex networked systems (e.g., the Internet) are emergent outcomes resulting from the interactions among individual entities (e.g., the autonomous systems in the Internet). Thus, the systems evolve in a bottom-up manner. In our previous studies, we have proposed a framework towards laying complex systems engineering on decision-centric foundations. In this paper, we apply that framework on modeling and analyzing the structure and performance of complex networked systems through the integration of random utility theory, continuum theory and percolation theory. Specifically, we propose a degree-based decision-centric (DBDC) network model based on random utility theory. We analyze the degree distribution and robustness of networks generated by the DBDC model using continuum theory and percolation theory, respectively. The results show that by controlling node-level preferences, the model is capable of generating a variety of network topologies. Further, the robustness of networks is observed to be highly sensitive to the nodes’ preferences to degree. The proposed decision-centric approach has two advantages: 1) it provides a more general model for modeling networked systems by considering node-level preferences, and 2) the model can be extended by including non-structural attributes of nodes. With the proposed approach, systems that are evolved in a bottom-up manner can be modeled to verify hypothesized evolution mechanisms. This helps in understanding the underlying principles governing systems evolution, which is crucial to the development of design and engineering strategies for complex networked systems.