| There are a large number of manufacturing processes in which parts must be mated with high precision. In particular, the automobile industry requires precise assembly of many parts. These include ball bearings, power steering shaft and sleeve assemblies, and fuel injector assemblies. The high quality of these parts is critical to the quality of the finished product. In such cases it is often either impossible or prohibitively expensive to produce parts so that every assembly satisfies the required specifications. Furthermore these high quality products are expected to be produced at a relatively low cost. To achieve these two goals (high precision and low cost) simultaneously, systems known as selective assembly processes have been developed. Despite their practical impact in industry, selective assembly operations have not received a lot of attention in the operations research and management science communities. In this dissertation we investigate various assembly methods for selective assembly operations on two problem types, so called simple matching problems and d-matching problems. We analyze quantitative and qualitative performance measures of each assembly method and computational experiments are performed to extend the analysis. Also probabilistic analysis is done to measure the effect of batch size in a mass production environment. Finally the multiple class assembly problem is addressed with mathematical programming approaches and several heuristic methods are suggested with discussion. |
