School of Computing and Information Sciences
Jerry Miller is a Ph.D. candidate at the School of Computing and Information Sciences, working on multi-variable, multi-dimensional network protocols under the supervision of Professors S.S. Iyengar and Niki Pissinou. He completed his B.S. in Basic Sciences at the United States Air Force Academy, a dual master’s degree in management and HR, and earned his M.S. in Telecommunications and Networking from FIU in 2011. He is a graduate of the Air Force Command and Staff College, Air War College, Defense Foreign Language Institute, and the Defense Institute of Security Assistance Management. He has published papers in several journals and conferences.
Multi-variable, multi-dimensional networks have historically used simplified 2-D network phenomena to model complex network variables, especially when these networks are expanded into 3-D topologies, resulting in a “not-quite-right” analysis. This has given rise to a plethora of routing protocols such as topology-based protocols, swarm-based protocols and geographic-based protocols working proactively, reactively or in a hybrid variation to provide efficient routing management. Yet, they do not accurately model the behaviors of networked systems by accounting for multiple interacting variables within the multi-dimensional spaces, nor do they account for the possibility of synergistic interactions which could create conditions of instability in the system resulting in a network crash and a corresponding condition of inactivity or a lower system performance equilibrium.
These multi-variable, multi-dimensional networks have not been fully explored mathematically or practically in how multiple variable forces or “affecting factors” relationally work upon the freely moving nodes within the network to cause shifts or failures. Mathematically, “affecting factors” can be summed into a single quantity, where the forces and object (node) will tend to move toward a position where the summation of forces is minimized. Analysis of these factors would enable better understanding of 3-D or 4-D network behavior.
As a first step, we propose to model these multivariable, multidimensional networks affected by sudden dramatic and continuous changes. We will do this by classifying the changes in behavior in the system through identification of the affecting factors which cause the system to undergo large changes in behavior, first establishing an NS-3 simulation baseline of existing network protocols most likely to be considered for use in multi-dimensional space. Next, we will apply mathematical theories most likely to identify the affecting factors and depict their impacts as curves and surfaces which can determine the critical point when a network state change occurs. By analyzing this information new network protocols can be developed and designed for greater efficiency and security in multi-dimensional networks.