Concurrent Machine and Batch Size Selection in Sampling-based Tolerance-Cost Optimization

Tolerance allocation is an essential part of detail design and an important interface between design and production. Its goal is to select tolerance values as tight as necessary, to reliably assure total product quality considering variation, but at the same time as wide as possible, to allow for a cost-optimal fabrication of the individual parts. A consequent front-loading of manufacturing information is thus inevitable for a realistic consideration of the given manufacturing conditions enabling an early, optimal process planning. In this context, sampling-based tolerance-cost optimization has proven its suitability as an optimization-based tolerance allocation technique because it allows a realistic consideration of part tolerance distributions as well as the selection of cost-minimal combinations of manufacturing machines and processes for each part. However, the potential of machine selection in tolerance-cost optimization is currently limited since it does not support distributed manufacturing on several machines with individual batch sizes. Motivated to close this gap, this article presents a novel optimization-based method for concurrent tolerance and machine allocation. The consideration of mixed batches from individually allocated machines requires the extension of the mathematical definition of the tolerance-cost optimization problem and its solution by advanced optimization and sampling strategies. Its exemplary application to a use case shows its potentials and proves that balancing tolerance values, machines and batch sizes can help to further increase the potential of tolerance-cost optimization. Hence, the presented approach supports a further shift to a manufacturing-oriented interpretation of tolerance allocation in design.