Flavors Technology Incorporated

Slide 18 of 24

The example shown above is the code required for an autonomous agent paint shop scheduler prototyped at GM Truck & Bus, Ft. Wayne, IN. The problem they were trying to address was to minimize the paint changes in a 10 line paint shop. Automotive paint can is very expensive, some colors reaching nearly $200 per gallon. Between colors, lines must be purged causing significant paint loss.

The autonomous agent scheduler created a scenario whereby each paint booth could bid on an incoming job. Bid currency was based on the color the booth was currently painting, and the space available in its queue. If a booth was painting the same color as the vehicle coming in, and it had room in its queue, its bid currency was higher. Taking this approach, paint savings were dramatic, projected to save almost $1M per year in unnecessary paint changes.