Scheduling Destructive Tests on Expensive Resources

Saturday, October 18, 2014 - 11:10am - 11:50am
Daniel Reich (Ford Motor Company)
Co-Authors: University of Michigan - Yuhui Shi, Amy Cohn, Marina Epelman; Ford Motor Company - Erica Klampfl

Ford’s Safety organization performs crash tests on prototype vehicles at multiple planning phases of each new vehicle program. These tests ensure the vehicles meet all government and company requirements by the time the vehicles reach the production phase. However, crash tests are quite expensive to perform due to the high cost of prototype vehicles compared with that of production vehicles. Accordingly, improvements in scheduling that reduce the number of prototypes crashed yield a significant cost savings. This scheduling problem has many sources of complexity: varying deadlines, precedence relationships between tests, incompatibilities in vehicle specs required, etc. Currently, engineers spend weeks of time manually planning the crash schedule for each new vehicle program and coordinating with all the other prototype vehicle users. We are developing an automated system for crash test planning that both minimizes the resources needed and significantly reduces the time engineers spend planning.