A Bayesian approach to sequential assembly experiments
Date
Authors
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Altmetrics
Authors
Research Projects
Organizational Units
Journal Issue
Is Version Of
Versions
Series
Department
Abstract
Sequential assembly experiments aim at identifying important sources of variability in performance attributable to component parts of an assembly with a few successive assembly tests. A test consists of a single- or multiple-part exchange on an assembly followed by unit reassembly and performance measurement. Two fundamentally different prescriptions for sequential experimentation with assemblies are developed. These are a Bayesian "look-ahead" heuristic and a "swapping" heuristic which is related to a technique currently used by practitioners;This study investigates and compares the characteristics of the heuristics through computer simulations. The context of this study is a linear random effects model for a three-part assembly with a single important source of variability. In addition, costs of experimental actions and a decision cost regarding the unknown identity of the important source are considered. The problem being addressed is determining whether one heuristic dominates the other in terms of criteria such as the success rate and total cost.