Predictive Modeling and Uncertainty Quantification with Application to Emergency Response

Problem Statement

We must be able to determine what is inside an unknown radiological system as quickly and accurately as possible. 

Technical Approach

A surrogate modeling approach based on polynomial chaos expansion is implemented into robust stochastic optimization/uncertainty quantification methods in an inverse transport toolset. 

Benefit

Surrogate modeling allows us to leverage the accuracy of stochastic approaches while achieving orders-of-magnitude reduction in run times.