Model Reification/Fusion Overview

This sub-session will detail the principals behind model reification/fusion, and will conclude with a demonstration of the model reification/fusion approach as implemented in BAREFOOT. First, the concept of reduced order models, truth models, and discrepancy models will be introduced, followed by the concept of reification (model correlation estimation). Finally, the model fusion as an optimization approach will demonstrated, highlighting its advantages.

Be knowledgeable of the terminology associated with model reification/fusion
• Truth Model
• Reduced Order Model
• Discrepancy Model
• Pearson Correlation Coefficient
• Perform simple model reification and model fusion
• Be knowledgeable of model reification/fusion as an optimization approach and how it fits within the BAREFOOT framework

Bayesian Optimization

This sub-session will cover the fundamentals of sequential Bayesian Optimization and will conclude with a demonstration of the Batch Bayesian Optimization approach as implemented in BAREFOOT. First, sequential Bayesian Optimization will be demonstrated, followed by Batch Bayesian Optimization; Both optimization schemes will be demonstrated on a simple ground truth model whose input and output are both 1-dimensional. The advantages of Batch Bayesian Optimization will be highlighted.

• Be knowledgeable of the fundamentals of Bayesian Optimization
• Be able to perform simple Bayesian Optimization and Batch Bayesian Optimization calculations
• Be knowledgeable of the advantages and disadvantages of both sequential and Batch Bayesian Optimization
• Be knowledgeable of how Batch Bayesian Optimization fits within the BAREFOOT framework

Multi-Objective Optimization Primer

This sub-session will serve as a brief primer on Multi-Objective Optimization, concluding with the construction of a Pareto-Front of optimal solutions.

• Be knowledgeable of the fundamentals of multi-objective optimization