Designing robust supply chains using a new hybrid push-pull control with multiple push-pull boundaries

The paradigm shift to 'mass-customization' from 'mass-production' has forced many organizations to change their supply chain design strategies from make-to-stock (MTS) to make-to-order (MTO). The impact of the trade-off between inventory cost and customer service level is critical in MTO supply chains. Push-pull strategy has been widely used to meet this challenge in many industries but it has the fundamental risk related to its order fulfillment capability, which gives rise to the major enhancement of its current methodology.;In the next stage of the research, the enhancement of analytical model and the identification of technologically feasible PPB locations by product, process and organizational form enable a comprehensive and integrated, high-level supply chain design approach that determines the strategic inventory point positioning in the context of semiconductor supply chain. A spreadsheet-based decision support tool is developed to provide supply chain managers with the capability to perform what-if analysis to evaluate supply chain design alternatives.;There are several key assumptions that limit the practical aspect of the analytical model for the proposed new push-pull strategy. In order to solidify the effectiveness of the model without those assumptions and to see the impact of the improvement on the 'push' portion of the hybrid push-pull model, a simulation optimization approach is used to compare the conventional push-pull strategy, the complete push strategy, and the proposed multi-point push-pull strategy. Numerical studies and experiments in all three different stages show that the proposed model produces more robust and cost-effective supply chain design alternatives against critical external factors in various input scenarios. Finally, the potential of the proposed model and the future research direction is discussed.;In this dissertation a new hybrid push-pull control model that can incorporate more than one PPB is presented as the improvement of conventional push-pull strategy. The new model leverages the concept of time and place postponement by placing customer orders to appropriate PPBs with sufficient probability of meeting target order lead times. Analytical framework of the model is developed mainly by the properties of the conditional expectation and variance with the optimization engine by Simulated Annealing (SA).