A Theoretical Framework for Understanding Strategic Behavior in Systems Engineering

Overview of the approach.

The objective of this project is to improve understanding about the impact of strategic behaviors under incomplete information on systems engineering projects. The complexity of modern engineering projects requires coordinating the activities of thousands of individuals across multiple organizations. Systems engineering tackles this complexity through well-documented processes for hierarchically decomposing the problem, passing requirements down the hierarchy, subcontracting, etc. However, such systems engineering processes typically result in significant cost overruns and schedule slippages. These represent significant losses to society and, oftentimes, taxpayers (one estimate values such losses to the U.S. Department of Defense alone at over $200 million per day). This research project will question assumptions underlying current systems engineering processes and advance understanding needed to improve systems engineering practice. To enable further research on this important problem area, research tools and data produced as part of this project will be made available freely via web-based repositories. A workshop will be organized to address research and educational challenges in this area. Results of this research can impact the development of large-scale engineered systems of any kind, including aerospace, defense, transportation and energy systems.

Sample results.

In this project, systems engineering processes will be modeled as dynamic network games of incomplete information. This will be accomplished in three modeling steps covering: interactions between engineers in system and sub-system levels, hierarchical systems engineering processes, and dynamic systems engineering processes. The overall expected outcomes of this project are (i) a foundational theory of systems engineering that considers the strategic decision making and distributed information, and (ii) a theoretical evaluation of existing systems engineering processes according to various criteria measuring their robustness to strategic behavior. The specific outcomes include information systems and reward/penalty structures encountered in established systems engineering processes, principal and agent utilities capable of taking into account empirical behavioral characteristics, characterization of agent design abilities and how they relate to the uncertainty of the process outcome, network game representation of established hierarchical systems engineering processes, characterization of hierarchical systems engineering network game equilibria, multi-stage network game representation of dynamic systems engineering processes, and characterization of dynamic systems engineering processes equilibria according to various criteria, e.g., the organizational utility at equilibrium, the total welfare of subsystem engineering teams, incentive compatibility, and more. The results of this project will be disseminated through publications, workshops, open source software, and an online data repository.



Salar Safarkhani
Ph.D. Candidate

Salar received his Bachelors degree in Mechanical Engineering from Amirkabir University of Technology (Tehran’s Polytechnic) in Iran. He received his Masters degree in Mechanical Engineering from Arizona State Univerity. He is currently a PhD candidate at the Predictive Science Lab at Purdue University.