ABSTRACT
We present a Decision Support System (DSS) which was developed at the IBM German Manufacturing Technology Center for the regular use in a number of production areas. These production areas are characterized by a job-shop-like environment consisting of machines with different technical capabilities (e.g. functions, speed, availability), and a varying demand involving 20–50 different part types. The planning horizon is in the range of 1–4 weeks. The problem is to determine a schedule such that the work can be completed within the planning horizon and the total processing time is minimized. The DSS is built around a set of mathematical optimization methods using mixed integer programming for scheduling problems. It has a menu driven user interface to adjust the generic model to the particular situation and to specify input parameters for the individual run. Results can be obtained textually (e.g. inform of tables) and graphically (e.g. in form of Gantt charts). Manufacturing supervisors use the system to confirm their production plan on a regular basis. Together with the required production volumes, they are able to specify release and due dates, part type priorities and current machine availabilities. As a result, they get an optimized production plan including the part-type-to-machine allocation, the parts loading sequence, and information on the individual machine utilization. Short response times allow the user to interact with the system in order to examine alternatives or to study the impact of parameter changes (e.g. machine availability) on the production center performance. Experience with the introduction and usage of the DSS in a printed circuit board manufacturing line will be reported.