IOWA STATE UNIVERSITY OF SCIENCE AND TECHNOLOGY
Intellectual Merit: The main objective of the proposed research is to uncover important mechanisms for the emergence of complex behavior in networked control systems, to propose ways to mitigate and control the effects of such behaviors, and to apply the findings to cooperative networked control applications.
Based on new and deeper understanding and the integration of information and control, the main idea is that the fading nature of the communication channels couples with the dynamics of the network nodes and can produce heavy-tailed distributions in the states of the networked systems. We argue that this coupling is tightly connected to (stochastic) uncertainty propagation in the system. One should therefore develop new analytical and computational tools for understanding and analyzing how uncertainty is generated, propagates, and can be managed in the system.
The novelty of our research is that we focus on networked systems to develop new foundations for complex systems. We do this for two main reasons:
1) Networked systems are the cradle for many complex behaviors and will become even more central in the future society.
2) They naturally bring the communication links to the forefront of our attention as the elements that reduce uncertainty (through information transfer) but that also introduce uncertainty in the system due to their unreliable nature.
At the beginning, we will focus on networks of linear systems with spatial invariant topologies. We will remove these assumptions after we gain some insights to study more general systems. Even with these restrictions though, the potential impact of the research is quite relevant since even networks of linear systems can show trademarks of complex behavior. Moreover, systems with spatial invariant topologies are a good approximation for large-scale systems and arise often in the study of partial differential equations, allowing us to bring a networked system insight into this important field.
The preliminary research by the PI, based on these ideas, has led to new results, which are very promising and indicate that vast advances can be made by this study. The PI intends to extend these preliminary results in several directions, described in detail in the proposal, that will directly impact numerous applications and disciplines. Important proposed directions are the identification of criticalities in the networked systems, the development of robust and adaptive approaches to manage complexity or criticalities while maintaining acceptable performance, and the development of systematic tools for network analysis and design.
Broader Impact: The deeper understanding and integration of control and information theory has opened the possibility for new progress in various fields involving complex systems. This new prospective makes this proposal unique and leads to great expectations for its outcomes. Because complex interconnected systems are becoming more widespread and dominant in today’s society, the proposed research program has the potential to greatly impact various disciplines, from control of electric power grid, to economic networks, from biological networks, to behavior of social networks.