Premium
A Systems Design Exploration Approach that Illuminates Tradespaces Using Statistical Experimental Designs
Author(s) -
MacCalman Alex D.,
Beery Paul T.,
Paulo Eugene P.
Publication year - 2016
Publication title -
systems engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.474
H-Index - 50
eISSN - 1520-6858
pISSN - 1098-1241
DOI - 10.1002/sys.21352
Subject(s) - computer science , set (abstract data type) , variety (cybernetics) , industrial engineering , statistical model , systems engineering , design of experiments , systems design , operations research , management science , simulation , machine learning , artificial intelligence , software engineering , engineering , mathematics , programming language , statistics
This paper describes an approach that leverages computer simulation models and statistical experimental designs for exploration studies during the early conceptual design of a system. We apply the approach to a naval ship design problem and demonstrate how we can illuminate trade decisions among multiple design decisions and evaluation measures using a dynamic dashboard. After performing experimental designs on a collection of simulation models, we can fit statistical models that act as surrogates to these simulations. These surrogate models allow us to explore a wider variety of system alternatives rather than fixating on a narrow set of alternatives. The purpose of the approach is to simultaneously explore the operational and physical domains using statistical surrogate models in order to illuminate trade decisions between the system's operational effectiveness and physical design considerations.