PITTSBURGH (August 13, 2018) … Relationships between return and investment cost, profit and time, or cost and quality are important for decision-makers looking to optimize efficiency. If the possible choices faced by the decision-maker have a simple structure, then these tradeoff problems can be solved efficiently; however, in practice, the decisions are rarely simple and the existing computational approaches fail after complexity reaches a certain point.The National Science Foundation (NSF) Division of Mathematical Science awarded $150,000 to Andrés Gómez, assistant professor of industrial engineering at Pitt’s Swanson School of Engineering, to widen the computational boundaries of complex optimization problems involving such tradeoffs. The project titled “Advancing Fractional Combinatorial Optimization: Computation and Applications” (1818700) begins Sept. 1.“We will be working with a hard class of problems called single- and multiple-ratio fractional combinatorial optimization problems,” Dr. Gómez explains. “There are no adequate approaches to these kinds of problems if they involve many layers of complexity or variability. This project aims to develop computational approaches with solid underlying theoretical foundations to solve these problems.”Dr. Gómez’s research falls broadly into the field of “decision-making under uncertainty.” He studies ways to improve mathematical modeling to better understand problems in finance, statistics, machine learning, manufacturing, revenue management, and many other applications.“Our proposed approaches will contribute to our understanding of mathematical optimization, particularly conic, fractional and discrete optimization, combinatorics, and algebraic graph theory,” adds Dr. Gómez.Oleg Prokopyev, professor of industrial engineering at Pitt, will join Dr. Gómez as co-principal investigator of the study.
Matt Cichowicz, Communications Writer, 8/13/2018
Contact: Matt Cichowicz