The complexities of our surrounding environments are becoming increasingly diverse, more integrated, and continuously more difficult to predict and characterize. These modeling complexities are ever more prevalent in System-of-Systems (SoS) simulations where computational times can surpass real-time and are often dictated by stochastic processes and non-continuous emergent behaviors. As the number of connections continue to increase in modeling environments and the number of external noise variables continue to multiply, these SoS simulations can no longer be explored with traditional means without significantly wasting computational resources.
This research develops and tests an adaptive sequential design of experiments to reduce the computational expense of exploring these complex design spaces. Prior to developing the algorithm, the defining statistical attributes of these spaces are researched and identified. Following this identification, various techniques capable of capturing these features are compared and an algorithm is synthesized. The final algorithm will be shown to improve the exploration of stochastic simulations over existing methods by increasing the global accuracy and computational speed, while reducing the number of simulations required to learn these spaces.
Identifer | oai:union.ndltd.org:GATECH/oai:smartech.gatech.edu:1853/44799 |
Date | 25 June 2012 |
Creators | Kernstine, Kemp H. |
Publisher | Georgia Institute of Technology |
Source Sets | Georgia Tech Electronic Thesis and Dissertation Archive |
Detected Language | English |
Type | Dissertation |
Page generated in 0.0015 seconds