AUTOMATIC PLANNING
Building schedules which use available resources in the most productive way is a difficult task because it requires envisioning a huge number of possibilities and finding the best one among them. For instance, there are typically billions of possible schedules for a group of 200 workers if each one should perform 10 tasks in a month; a human planner simply cannot review them all in order to find the best possible one.
Mathematical methods such as linear or integer programming can tremendously help to build productive schedules. However, Resource Management problems often turn out to be so large that general optimization techniques are too slow: they take days or even weeks or months in order to find an optimal solution.
This is where one of Kernel Software’s strengths lies: not only is its team of optimization specialists proficient in general optimization techniques, it also knows what to do when such techniques need to be combined with domain-specific knowledge. The Automatic Crew Assignment module that Kernel Software provides Lufthansa with, demonstrates this point: month after month, it builds the monthly schedules of Lufthansa’s 5,000 pilots and 25,000 flight attendants. A task which used to keep many planners busy for two weeks is now performed in a few hours, with better results and with fewer planners.
