Abstract
In the semiconductor industry, assembly is the important stage in backend operation. To focus on core competence and reduce capital expenditure, many semiconductor companies adopt assembly outsourcing for reducing the manufacturing cost and business risk. Thus, assembly outsourcing is a critical component involved in the semiconductor supply chains. Furthermore, semiconductor companies usually maintain multiple assembly vendors for reducing assembly time and production risk. Since lead time of assembly materials is about one or two months, monthly material requirement planning will be determine to allocate orders to the vendors. Based on monthly material requirement planning, daily real time order assignment planning is then needed to assign the orders to specific vendors in real time. Daily order assignment planning considers the multiple objectives and constraints based on real time information. It is a complex real time decision process on daily order assignment in semiconductor assembly outsourcing. To address this need in real settings, this study aims to construct mixed integer linear programming (MILP) and goal programming (GP) model for real time order assignment problem and develop decision support system to implement in semiconductor assembly outsourcing. The proposed model not only considers the general objectives involved in production planning but also considers the shuttle status and huge lot issues. Since this problem is a real time decision process, we apply technology of data preprocessing to reduce the amount of data for reducing solving time before solving this problem. We conducted an empirical study in a semiconductor company for validation. The results showed the solving speed of proposed model is more efficient than heuristic method based on expert experiences and the results can also achieve the objectives under the limitations of constraints in real settings. Furthermore, we developed a daily assignment model and decision support system embedded with the proposed model to provide associated management functions to assist the decision makers in the semiconductor companies. Future research can be done to analyze real time assignment problem from vendor’s perspective.