Sequential Testing for Complex Systems and Cost Structures

Modern manufacturing systems, infrastructure and healthcare technologies rely on testing and inspection to ensure that critical components function safely and reliably. Because these systems are becoming increasingly complex, determining which tests to perform, and in what order, is an important challenge that directly affects cost, efficiency, and decision making. This award supports research that aims to develop novel optimization methods to design testing strategies that obtain reliable information while minimizing the time and resources required for inspection. The resulting advances have the potential to improve testing and quality assurance in a wide range of engineered systems, contributing to more efficient manufacturing, more reliable infrastructure, and better-informed operational decisions. The project will also provide research opportunities for graduate and undergraduate students in optimization and data-driven decision making, while increasing participation through workshops and outreach activities. This project will develop new optimization and analytical tools for sequential testing problems under complex cost structures. The research investigates the fundamental tradeoff between performing multiple tests together to reduce costs and performing individual tests to acquire information more rapidly. Building on recent advances in approximation algorithms, the project will develop efficient algorithms with mathematically rigorous performance guarantees for testing complex systems composed of interconnected components, including series-parallel structures. These methods will be evaluated using publicly available datasets and are expected to advance the theory of sequential testing under realistic cost structures. More broadly, the resulting techniques are anticipated to contribute to the design of efficient solution methods for stochastic dynamic programs with exponential state space, where exact computational methods are intractable. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.