Dr. Alejandro Salado earns NSF award on building a scientific theory of verification
(July 17, 2018) Assistant Professor Dr. Alejandro Salado has been awarded by the National Science Foundation (NSF) for his work on value-maximizing system verification strategies in multi-firm engineering projects. Dr. Salado will collaborate with a former VT ISE faculty member, Dr. Christian Wernz at Virginia Commonwealth University, on the project, “Collaborative research: maximizing the expected value of engineered systems through coordinated verification.” With approximately $400,000 in grant money over a three-year period, the pair of researchers will develop a mathematical framework of incentives that will enable engineering teams to define their own verification strategies while optimizing the value of the overall system.
For more information about the project, please read the following project abstract.
Maximizing the expected value of engineered systems through coordinated verification
Verification shapes the confidence engineers have on the proper functioning of a system. Because resources are limited, absolute certainty is not feasible and engineers need to decide which verification activities to carry out, at which integration level, and with which accuracy. Although the aggregated verification effort of all engineering teams should contribute to maximizing the expected value of the overall system, engineering teams seek to optimize their individual objectives instead. This conflict of interest yields verification strategies that are driven by post-negotiations and risky de-scoping actions when under schedule and budget pressure. This research project will develop a mathematical framework of incentives as a coordination mechanism that will enable engineering teams to define their own verification strategies, while maximizing the value of the overall system. If this research is successful, the public will benefit from increased safety and efficacy of commercial products and public services. Moreover, the systems engineering and engineering design research communities will benefit from a foundational, mathematical framework, which enables a consistent integration of verification strategies and contractual structures upon to which build future research. In addition, this research will advance the interests and skills of women and other underrepresented groups by actively recruiting and mentoring them as part of this project.
The objective of this research is to mathematically derive value-maximizing system verification strategies in multi-firm engineering projects. The proposed method enables computing verification strategies that maximize a firm's objective function, while accounting for the interdependencies between the firms, the components they develop, and the value of the overall system. Moreover, the framework will account for and enable the design of contractual agreements between firms in order to ensure that locally optimal firm verification decisions contribute to the global goal of maximizing system value. To achieve this research objective, mathematical methods from different disciplines - operations research, decision theory, and economics – will be combined and advanced into one framework. Specifically, partially observable Markov decision process will be employed to derive optimal verification strategies for each firm, multiscale decision theory to mathematically capture the interdependencies between firms, and contract theory to design effective contracts that can implement globally optimal verifications strategies. This work will contribute to build a scientific theory of verification.