Changing Subsystem Information Strategies Using Weighted Objectives: Increasing Robustness to Biased Information Passing

• Main Authors: Austin-Breneman, Jesse (Contributor), Yu, Bo Yang (Contributor), Yang, Maria (Contributor)
• Other Authors: Massachusetts Institute of Technology. Department of Mechanical Engineering (Contributor), Massachusetts Institute of Technology. Department of Ocean Engineering (Contributor), Massachusetts Institute of Technology. Institute for Data, Systems, and Society (Contributor), MIT Edgerton Center (Contributor)
• Format: Article
• Language: English
• Published: ASME International, 2019-01-14T18:26:11Z.
• Subjects: Article
• Online Access: Get fulltext

Complex system design requires managing competing objectives between many subsystems. Previous field research has demonstrated that subsystem designers may use biased information passing as a negotiation tactic and thereby reach sub-optimal system-level results due to local optimization behavior. One strategy to combat the focus on local optimization is an incentive structure that promotes system-level optimization. This paper presents a new subsystem incentive structure based on Multi-disciplinary Optimization (MDO) techniques for improving robustness of the design process to such biased information passing strategies. Results from simulations of different utility functions for a test suite of multi-objective problems quantify the system robustness to biased information passing strategies. Results show that incentivizing subsystems with this new weighted structure may decrease the error resulting from biased information passing.
University of Alabama in Huntsville. System Engineering Consortium
National Science Foundation (U.S.). Graduate Research Fellowship