A scheduling heuristic for maximizing wirebonder throughput

Process planning and scheduling is an important part of managing the complex dynamic back end (i.e., test, assembly, and packaging) of the semiconductor manufacturing process. One of the most critical toolsets in the back end in terms of manufacturing capacity and product flow is the wirebonders that electrically connect bare silicon die to lead frames via thin metal wires. We develop two approaches for maximizing throughput on a group of wirebonders containing three different wirebonder types. Although our mathematical programming approach is capable of producing a near-optimal wirebonder schedule for any production day of interest, the computational requirements of this approach are unacceptable for implementation in a real-world back-end facility. Therefore, we develop a rule-based heuristic approach for wirebonder scheduling capable of producing high-quality schedules in a much shorter period of time.