| The current trend in industry is that lighter and smaller products are becoming more favorable. However, at the same time, design and manufacturing processes for such products become more complex and it causes an increasing demand for sophisticated, computer mediated design tools to aid designers during product development and manufacturing. In this study, a computational methodology has been developed for packaging optimization during the product design process.; Packaging optimization is categorized into three processes: preliminary packaging generation, constraint satisfactions, and packaging verification. At preliminary packaging generation process, candidate packaging is generated. We have developed three different component placement schemes, namely, position-based method, sequence-based method, and combined sequence and position method. Position-based method generates a packaging by positioning all components at the same time. On the other hand, sequence-based method packs one component at a time in a sequential order. Compare to position-based method, sequence-based method is faster but has a higher tendency of being trapped in local optimum. Combined sequence and position method is based on sequence-based method but it allows random positioning of components to prevent trapping in local optimum.; The approach uses mating between a pair of components as means to accelerate the packaging process. In this study, mating between two components has been defined as positioning one component relative to another by merging common features that are assigned through mating conditions between them. A constrained move set is derived from the mating condition that allows the transformation of a component in each mating pair to be fully or partially constrained with respect to the other. By using the mating in the packaging, thus, the number of components to be placed can be reduced significantly and overall speed of the packaging process can also be improved.; Candidate packaging is then checked for their constraint satisfactions . In this study, a new and efficient methodology of detecting an overlap has been developed. This method introduces the concept of Minkowski sum and by using simple mathematical relations, it can rapidly detect if two components are overlapping. Packaging verification is the final phase of the packaging optimization process in which the packaging generated at the previous stage is refined through more detailed performance analysis and evaluation. Quantitative results of packaging goals along with penalties are added and then the design is rated according to the results. In the meantime, genetic algorithm is used to drive the design improvement based on the quantitative results. |
