You are here
Chaos In Distributed Systems?
Phone: (703) 759-5257
Is chaos present in any significant real distributed system? We don't know. We will use our commercial grade toolkit for chaotic data analysis to see if chaos is present is some distributed systems. Our toolkit incorporates all the relevant methods for detecting and analyzing chaos. The remarkable aspect ofthese methods is that they operate directly on data. No prior knowledge or model of the system is required. We can directly determine the degrees-of-freedom (dimensionality) of the system and the rate of which information about system state is lost for all degress of freedom. Our collaborators at the Institute for Nonlinear Science, UCSD, invented many of the techniques, including local statistics. Local statistics examine how systems evolve in finite time (vice the entire data set). Local dimensions are important for real world prediction, control, and modeling. We also present several methods for controlling chaos. Anticipated Benefits: These methods may provide better ways of controlling large distributed computing environments, logistics networks, C4I networks, and transportation systems.
* Information listed above is at the time of submission. *