BYU & ATK Thiokol Team-up in Research

December 31, 2006

Research with industrial partner company ATK was highlighted in the Computer Science news-eletter, CS Connections:

Dr. Sean Wamick and two undergraduate students recently completed a project with ATK Thiokol, the world’s leading supplier of rocket motors for space launch vehicles, strategic missiles, and missile defense interceptors. ATK Thiokol is headquartered in Brigham City, Utah and is part of ATK with headquarters in Minnesota. This branch of ATK has annual sales of $845 million dollars.

Dr. Warnick was sought out by the company and asked to create a mathematical model to optimize the production of various propellants for a new factory design. David West, an undergraduate student, worked with Dr. Warnick in the beginning of the project and researched manufacturing systems and how they operate. The factory would need to produce different kinds of propellant with the same capital equipment and different “recipes” for each propellant type. Given an annual quota for the various propellants, the question was, what is the minimal amount of time necessary to produce the quota and in what order should they manufacture the propellants to realize the minimal time?

This is a classic scheduling problem, but some peculiar features of the ATK factory made the problem even more interesting. For example, propellants are in batches and the machines process different sized batches of different propellant types, and since propellants can explode, no inter-machine storage or stockpiling is allowed. This implies that although a batch of propellant may be finished processing in one machine, it can not unload if the next machine is still occupied. Thus “bottlenecking” can be a major problem.

After developing a mathematical model of this type of manufacturing problem, the team began to formulate a Traveling Sales Person (TSP) approach to a solution. At this time Sam Weyerman, another undergraduate student, joined the team and developed a simulation of the factory and a TSP-like formulation of the problem as an integer program.

Weyerman computed a solution that is approximately 15% better than the ATK team’s current best solution. The factory, if scheduled properly to avoid bottlenecks, can produce 15% more propellant of the various types per year than was previously thought possible. The approach taken by Weyerman to compute effective solutions to these types of scheduling problems is currently being submitted for publication. Weyerman will be a graduate student working with Dr. Warnick in January.

ATK Thiokol is one of the Industrial Partners sponsoring the Information Dynamics Group, which is led by Dr. Warnick. Other partners include the BYU Bookstore, the Utah Bureau of Reclamation, and Sandia National Laboratories. Amazon.com and Utah Power are currently discussing project ideas with Dr. Warnick’s Lab.

Although these partners often need specific answers to questions, the approach of this Industrial Partners Program is different from consulting. Dr. Warnick’s Industrial Partners sponsor student researchers to explore a general area of research of interest to the partner, such as Batch Scheduling Problems, Demand Forecasting, or Dynamic Pricing. These student researchers then draw from their expertise with mathematics and algorithms design to precisely formulate the problem and develop a computable solution or design method for the general problem. The specific questions an Industrial Partner may have then become special cases of the new tools developed through the group’s research efforts.

Dr. Warnick strongly advocates this approach to practical problems because it develops something more important than answers to today’s questions. Students learn to become technical problem-solvers, capable of developing precise problem formulations and efficient solutions for tomorrow’s problems, independent of any particular application area such as manufacturing or eCommerce. “Industry has so many challenges,” he argues, “that problem solvers are much more valuable to our partners than a handful of solved problems.”


Read More