EQUiPS: Enabling Quantification of Uncertainty in Physical Systems

Complex physical systems, devices and processes important to the Department of Defense (DoD) are often poorly understood due to uncertainty in models, parameters, operating environments and measurements. The goal of DARPA’s Enabling Quantification of Uncertainty in Physical Systems (EQUiPS) program is to provide a rigorous mathematical framework and advanced tools for propagating and managing uncertainty in the modeling and design of complex physical and engineering systems. Of particular interest to the program are systems with multi-scale coupled physics and uncertain parameters in extremely high-dimensional spaces, such as new aerospace vehicles and engines. Despite considerable progress in propagating input uncertainties through forward models, current approaches have limitations that prevent their application to many important DoD and industrial systems. For example, Uncertainty Quantification (UQ) methodologies do not scale to large, multi-scale systems, and they do not directly address the challenges of UQ in inverse problems, which include estimation of unknown parameters from noisy data, and model uncertainty. Novel mathematical research is needed for dealing with the underlying high dimensionality of the space of uncertain parameters, strong multi-physics coupling, and uncertainty in the models themselves. In addition, there is no fundamental mathematical theory for decision making and design under uncertainty for these large-scale dynamic systems. To address these shortcomings, EQUiPS intends to develop the following capabilities:

• New methods for forward and inverse modeling to scale to high-dimensional multi-scale/multi-physics systems
• A quantitative understanding of uncertainties and inadequacies in the physical models themselves
• A completely new paradigm for stochastic design and decision making for complex systems