In the automotive industry, ever-shorter product life cycles and growing product portfolios have led to a high frequency of new product launches. The production systems for these products are usually very complex and product specific. Therefore, each new product introduction is associated with a project to engineer and implement its production system. For undertaking these projects, car manufacturers deploy inhouse manufacturing engineers and use outsourcing services from engineering services providers. Planning these projects requires the car manufacturer to decide on how many inhouse engineers to deploy, how to assign these engineers to activities, and which activities to outsource to which service providers. In practice, this planning problem is solved manually and is not optimized. Therefore, we present a multi-objective mixed integer programming model based on the resource-constraint multi-project scheduling problem to solve the planning problem. We consider two lexicographic objectives. The first objective maximizes the utilization of inhouse engineers, which is equivalent to minimize outsourcing. The second objective minimizes jumps in the resource profiles for the service providers to generate attractive outsourcing bundles. Currently, we are developing a decomposition algorithm based on column generation to solve the problem. The algorithm exploits the block structure of the problem and allows to generate schedules for each project independently. Furthermore, we are currently collecting data from a major European car manufacturer to conduct a computational study to obtain managerial insights into optimal project schedules, workforce compositions, and outsourcing strategies.
AdONE Seminar: Max Kolter (AdONE, TUM)