Product variety induced complexity and its impact on mixed-model assembly systems and supply chains

Product variety has increased explosively in the past few decades and brought many challenges to manufacturing systems and supply chains. This dissertation studies the product variety induced complexity and its impact on mixed-model assembly systems and supply chains.;The first part of this dissertation develops a complexity measure for mixed-model assembly systems with different configurations, including serial, parallel and hybrid configurations. The impact of complexity on system throughput is analyzed using an approximate throughput model, which takes into consideration the operator reaction time and fatigue effect. The complexity and throughput models are then used to compare the performances of assembly systems with different configurations. Finally, a mathematical formulation is developed based on mixed-binary nonlinear programming to minimize the complexity or maximize the throughput of a mixed-model assembly system by allocating the modules to different stations.;The second part of the dissertation develops a complexity measure for assembly supply chains based on the information entropy. This complexity measure takes into account the supply chain configuration, the variety level at each node in the supply chain, and the demand ratios of the variants offered by the node. In addition, the degree of consistency between the complexity and cost is studied when the complexity and cost are used to compare (1) assembly supply chains with the same configuration but different levels of product variety, and (2) assembly supply chains with the same level of product variety but different configurations.;The third part of the dissertation applies the complexity of assembly supply chains to configuration design, i.e., finding the optimal supply chain configuration with minimum complexity given the product variety level at the final assembler. The optimal supply chain configuration is studied for two special scenarios: (1) the demand share of one particular variant is bigger than that of others at the final assembler, and (2) demand shares are equal across all variants at the final assembler. In addition, a methodology is developed to find the optimal supply chain with/without assembly sequence constraints for general demands.