Semiconductor supply chains are challenged by an intense international competition, technological complexity, and a high innovation rate, which is typical for the semiconductor industry. This dissertation addresses these challenges by proposing optimization-based approaches for three different decision problems in the areas product platform design and hierarchical production planning. The proposed approaches are applied to cases from the semiconductor industry and extensive numerical experiments are conducted for validation and evaluation. First, stochastic optimization is introduced for product platform design in silicon wafer manufacturing. Numerical results show that taking the uncertainty about future demand explicitly into account helps to design product platforms optimally – also to the requirements of future customer orders – and thus reduce future design workload and costs. Second, a novel cycle time-oriented mid-term production planning model is applied to wafer fabrication. Tightly integrated with production control, the model ensures an optimal response to machine failures and unforeseen demand changes. Compared to conventional work-in-process-oriented planning, cycle time-oriented planning delivers higher service levels, shorter cycle times, and it generates simpler production plans. Third, a low-dimensional capacity model is suggested for company-wide production planning. It hides the detailed capacity allocation decisions, which are usually made for parallel machines, and thus reduces the complexity of the planning process. Depending on the number of modeled machines and products, an exact or a heuristic procedure is used for the generation of specific capacity constraints. An aggregation step, which exploits certain attributes of machines and products, reduces the problem size and computation time. Compared to existing methods, the proposed procedures deliver a more accurate representation of throughput limitations, in particular for parallel machines. The presented optimization-based approaches improve the resource efficiency and the service level of semiconductor supply chains and thus strengthen the resilience of manufacturers against future challenges.
«
Semiconductor supply chains are challenged by an intense international competition, technological complexity, and a high innovation rate, which is typical for the semiconductor industry. This dissertation addresses these challenges by proposing optimization-based approaches for three different decision problems in the areas product platform design and hierarchical production planning. The proposed approaches are applied to cases from the semiconductor industry and extensive numerical experimen...
»