Seminar | Our Top Applications of IEOR Past & Present: Truck Routing, Voting Lines, Hospital Alarms, and More

Seminar by Theodore T. Allen, PhD

Associate Professor

Department of Integrated Systems Engineering

The Ohio State University

This presentation provides a broad survey of projects that involve operations research, human factors, cybersecurity, and machine learning under a translational data analytics paradigm. The literature and innovations include: (1) Saving over $87M through near-optimal routing with DHL; this was achieved in part with methodological advances in "red-black" ant colony search and dynamic programming (Informs Journal of Applied Analytics), (2) Helping millions of citizens to have shorter voting lines across more than 5 states using in part generalized binary searches with indifference zones (IISE Transactions, POMS, & CAIE), (3) Achieving 17-59% reductions in response times for Code Blue events in four hospitals by a redesign of the alarm tones and eliminating some redundant alarms (Rayo); pre-post differences were analyzed with staged control charts (Human Factors), (4) Reducing hospital discharge times at an Ohio hospital unit by 36 minutes using lean six sigma (Quality Engineering), and (5) Increasing printed circuit board yields by 30% using experimental design at a local Ohio manufacturer (Intro. To Engineering Statistics…3rd Edition).

Ongoing top five projects are also described including: (1) Seeking to reduce cybersecurity maintenance
and incident costs by over 50% using social media analytics, optimal classification trees, and partially observable Markov decision processes (NSF, Ohio Dept. of Higher Edu., SecureAmerica), (2) Decreasing inspection costs and oil and gas pipeline leaks, we seek over 99.5% corrosion classifications using "pure leaf" trees and classification with the rejection option (Rosen Group), (3) Increasing vaccination rates through targeted social media advertising and experimental design (FactSpread), (4) Improving throughput and inventory using "human-led-optimization assisted" scheduling (Walsin Sinha), and (5) Battlespace Optimization with Tabu Efficient Global Optimization (TEGO, Infinity Labs-AFRL).

Theodore T. Allen, Ph.D. is an associate professor of Industrial and Systems Engineering at The Ohio State University. He is the author of over 60 refereed publications focusing on the intersection of optimization and experimentation. These include two textbooks published by Springer Verlag including a third edition textbook on engineering statistics. Dr. Allen is a fellow of ASQ. He is an area editor of Computers and Industrial Engineering (IF:3.5) and a past associate editor of the Journal of Manufacturing Systems (IF:3.6) and has been a reviewer and on the editorial board for numerous journals and organizations including Technometrics and the AIAA journal. Dr. Allen was a co-author on apparently the first optimal design paper for multi-fidelity experiments (Huang and Allen 2005), the first extension of efficient global optimization to both noisy and multi-fidelity cases (Huang, Allen, Notz, and Zheng 2006 – highly cited paper and Huang, Allen, Notz, and Miller 2006 – highly cited paper), and the first application of multi-fidelity optimization in the context of discrete event simulation (Schenk, Zheng, and Allen 2005). Dr. Allen is also the president and founder of the educational non-profit organization factSpread.