Near Real-Time Quantification of Stochastic Model Parameters

Abstract

We quantify the uncertainty in estimated model parameters using both a Bayesian and a Frequentist approach. We then apply these methods to a class of quasi-chemical models (QCM)developed by the U.S. Army Natick Soldier Research Development and Engineering Center(NSRDEC) (References [a] and [b]). The QCM models, developed to model bacteria growth on food under various environmental conditions, are capable of capturing the effects of microbial lag, inactivation and tailing. Bayesian and Frequentist approaches for solving the inverse problem are presented in this report. Solutions using the two approaches for the datasets from Reference [b] are compared. Uncertainty Quantification (UQ) methods for Forward Uncertainty Propagation are also applied using the datasets.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 28, 2016
Accession Number
AD1100945

Entities

People

  • H. Thomas Banks
  • Jared Catenacci
  • Panagiotis Tsilifis
  • Paul K. Newton
  • Roger Ghanem
  • Thomas E Wood
  • William J. Browning

Tags

Communities of Interest

  • Autonomy
  • C4I
  • Space

DTIC Thesaurus Topics

  • Algorithms
  • Applied Mathematics
  • Bacteria
  • Bayesian Networks
  • Birds
  • Chemical Kinetics
  • Computational Science
  • Computer Programs
  • Data Sets
  • Design Criteria
  • Differential Equations
  • Equations
  • Experimental Design
  • Gaussian Distributions
  • Inverse Problems
  • Machine Learning
  • Measurement
  • Microbiology
  • Monte Carlo Method
  • Normal Distribution
  • Probability
  • Probability Distributions
  • Random Variables
  • Test Methods

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Computational Modeling and Simulation
  • Microbial Pathology

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • Biotechnology