Making better choices, from power to payloads

By John Sullivan
January 04, 2018

This article is from the Winter 2018 EQuad News.

Warren Powell ’77, a professor of operations research and financial engineering, has developed methods to help make complex choices in fields from shipping to power management. (Photo by David Kelly Crow)

Many scientists’ work begins with a question in the lab. Warren Powell ’77 had a different start.

“My career began with a dinner conversation,” said Powell, a professor of operations research and financial engineering.

At the dinner, in 1982, a trucking executive complained that his company’s system for matching loads and drivers assumed they could predict the future.

The frustrated executive, with a master’s in operations research, stated that, in reality, the problem was “stochastic,” meaning inherently unpredictable. Powell saw this as a challenge.

“I have this paper from the 1980s; it has four completely different ways to model the stochastic fleet management problem,” Powell said. “All of them bad.” Powell worked at the problem for the next 20 years, applying versions to railroads and massive electrical grids. “I kept chipping away at it,” he said.

Thirty years, 230 papers, and two books later, Powell has applied his system in many fields, including blood management, spare parts supplies, medical decisions, and power grids balancing the uncertainties of wind and solar energy. Powell said he is still working on the basic questions. “Even in the last year or so, I have ‘aha’ moments when I am teaching my grad courses,” he said.

His team currently has a new model in production — it matches truck drivers to loads.