IOE-STAGE SITE
The grant will fund the development of cutting-edge algorithms for solving complex nonlinear constrained stochastic optimization problems. Albert Berahas, and Micheal O’Neill will lead the research to advance optimization methods and address real-world challenges.
Albert S. Berahas an Assistant Professor at the University of Michigan’s Industrial and Operations Engineering (IOE) department, has been awarded a grant from the Office of Naval Research (ONR) to explore advanced algorithms for solving nonlinear constrained stochastic optimization problems.
The grant, titled “Advanced Algorithms for Nonlinear Constrained Stochastic Optimization,” will support the design, analysis, and implementation of cutting-edge sequential quadratic programming (SQP) and interior point (IP) methods tailored to these complex problems.
Optimization problems are at the core of many scientific and engineering applications, ranging from machine learning, healthcare, power systems, statistics and much more. Often, these problems involve nonlinear, nonconvex, and stochastic elements, making them particularly difficult to solve efficiently. This project aims to bridge this gap by developing robust and adaptive algorithms that handle real-world constraints and stochastic behaviors.
“This project has the potential to significantly advance current methodologies for solving stochastic optimization problems involving general, nonlinear and deterministic constraints,” said Berahas. “Such problems arise in a plethora of applications. We expect this collaboration with the ONR will lead to new algorithms that will allow researchers and practitioners to solve challenging optimization problems that are beyond the reach of current state-of-the-art methods.”
The collaboration includes co-principal investigator Michael O’Neill from the University of North Carolina’s Department of Statistics and Operations Research. Together, Berahas and O’Neill will work to develop the algorithms and validate their performance on engineering design and machine learning applications.
The proposed work aligns with ONR’s focus on mathematical innovation and resource optimization, particularly under their Mathematics, Computer and Information Sciences (MCIS) Division. By leveraging advanced features such as adaptive step sizes, second-order information, and robust parameter estimation, the research aims to push the boundaries of optimization capabilities, particularly in defense-related applications.
This project not only opens new doors for tackling optimization challenges but also strengthens the collaboration between experts in engineering design and machine learning, ensuring the results of this research have a broad and lasting impact on multiple fields.
Photo caption (top): Assistant Professor Albert S. Berahas teaching IOE 511 at the Mortimer E. Cooley Building in Ann Arbor, Mich.