Explore the field of Bi-Fidelity Expensive Black-Box (Bf-EBB) problems in this 29-minute seminar from GERAD Research Center. Delve into the complexities of multi-fidelity optimization, focusing on scenarios with expensive yet accurate information sources alongside cheaper, less precise alternatives. Examine three variations of Bf-EBB problems through a toy problem illustration. Investigate the reliability of surrogate model techniques in solving these challenges, particularly in single low-fidelity function contexts. Consider the crucial question of how much to rely on low-fidelity data sources when tackling Bf-EBB problems. Gain insights into fidelity surrogate modeling and its applications in optimizing design outcomes and decision-making processes.
Overview
Syllabus
Introduction
Definitions
Australia
Sub Problem 1
Sub Problem 2
Two types of analysis
Integrate low Fidelity data
Illustration
BiFidelity
Examples
Explanation
Fidelity surrogate modeling
Taught by
GERAD Research Center