History

Though the Department is relatively new, Operations Research & Financial Engineering has had a distinguished history at Princeton. In fact, a significant portion of modern ORFE is an outgrowth of activities at Princeton between 1930 and 1960.

Princeton faculty played a major role in the development of the theory of convex optimization (both as it relates to individuals and groups). Indeed, H. Kuhn and A.W. Tucker (in the 1950s) and R.T. Rockafellar (in the 1960s) played an important role in the development of much of convex analysis. In addition, R. Gomorry laid the foundations of integer programming in the 1950s, and J. Nash, J. von Neumann, and O. Morgenstern (in the 1930s, 40s and 50s) did seminal work in the area of multi-agent, non-cooperative optimization (i.e., game theory).

Modern probability and statistics have also been influenced significantly by members of the Princeton faculty. S. Wilks (in the 1930s) and W. Feller and J. Tukey (in the 1950s) all played major roles, and G. Hunt developed the potential theory of Markov processes in the early 1960s.

And, of course, the use of digital computers in operations research owes much to Princeton. A. Turing and J. von Neumann (in the 1930s) and A. Church (in the 1950s) made enormous contributions in such areas as the theory of cellular automata, computer memory, computer reliability, and the lambda-calculus.

This distinguished tradition continues today as Princeton continues to lead the way. In fact, Princeton is the first major university to see the importance of ORFE and create a department devoted to its study and teaching.

Present

Research and teaching in the Department of Operations Research and Financial Engineering (ORFE) focuses on the foundations of data science, probabilistic modeling, and optimal decision- making under uncertainty. Many aspects of society and the workplace are being transformed by the ongoing information and machine learning revolution. The concepts and tools for making strategic decisions in many fields have their origins in advances from ORFE subjects. The areas where these skills are useful include communication systems and services, economics and finance, energy and environmental science, healthcare management, physical and biological sciences, as well as transportation services and networks.

The graduate and undergraduate core ORFE programs are popular since they provide students with a rigorous technical grounding in a knowledge discovery methodology. These techniques transform data, records of what is happening, into informed decisions based on an analytical sense of what should happen. The data science subject of statistics studies the derivation of information or patterns within the data. This allows us summarize what was happening. An information rich description summary of the original system yields a model that tells us what does happen. It is the analysis of these models that tells us what might happen. The randomness of uncertainty, that we should ignore, is typically called noise. However, the randomness of uncertainty, that we cannot ignore, is referred to as risk. The probabilistic subject of stochastics studies the relevant random dynamic models for an ever changing world of uncertainty filled incomplete data and risk. Understanding the consequences of various actions leads to a search for the sequences of actions that yield the best outcome. The subject of optimization generates through algorithms the right sequence of actions for the best policy of what should happen.

Recently, the primary driver of major breakthroughs for statistics, stochastics, and optimization has been their own collective synergy. Novel insights into large-scale optimization and high- dimensional random structures is one synergistic example. They play an increasingly crucial role in the statistical analysis of massive datasets ranging from the human genome to financial data. Moreover, ideas from statistics and machine learning are now the central tools for the development of new decision making algorithms.