Modelling of information sharing in capacitated assembly supply chains using simulation

It has been noted in Lee and Wu (2006) that competition has required the integration of supply chain members and resources through supply chain management (SCM) and that inventory control as a key component of SCM is no exception to this. One manner in which firms in a supply chain can integrate is through information sharing. New technologies have facilitated the movement and processing of data between partners such as electronic data interchange (EDI) in the past, through to more current tools such as on-time, real-time web based portals. In tandem with more readily available information and data, methods for utilisation of such have been explored in research and practice. An array of literature now exists on various aspects of information enabled supply chain management and information enabled inventory control. Considering this, we have proposed a dissertation in the field of information sharing in capacitated assembly supply chains focusing on the inventory control mechanisms applied within. A case study based around a contract manufacturer in the electronics industry provides a vehicle for testing the information sharing propositions within a realistic supply chain. It is complicated on the supply side by multiple suppliers, constrained capacity and varied delivery lead times, and on the demand side by stochastic periodic demand and unpredictable short-lived upturns or spikes in end item demand. The case study partner also experiences the complications of being the convergent partner/node in an assembly system. It orders components from nine independent suppliers to be combined into a single product which is assembled for an OEM. It is typical of many such companies in the multibillion euro electronics manufacturing services industry. The system is modelled in a simulation tool and assessed through a range of metrics, across an array of experimental control parameters all derived from the case study and literature. Important insights to the dynamics of the system are commented upon, and quantitative analysis is provided of the performance of the information enabled inventory control policies. Rather than being a panacea where ubiquitous information is simply available to all partners, successful information sharing to support supply chain management requires two key elements: (1) mechanisms of use of different types of information applied to change inventory policies, and knowledge of how these mechanisms interact with each other and (2) thorough analysis of the prevailing supply chain conditions as these influence the effectiveness of the mechanisms used. This contribution of this thesis is a comparison of information sharing options to inventory replenishment including a novel application of point of sale data, and the management of a buffer at the converging node based on lead-time and capacity information. It is shown that tailoring information sharing to the conditions, especially uncertainty, in supply-chain demand yields better performance. Demand characterisation was indicated as a key determinant in performance ranking between the point of sale mechanism and the periodic review policy used as control: it is not adequate to model spiked demand by simply increasing variability. Information-determined buffers outperformed flat buffers under certain supply chain conditions, but not always as a rule. It was found that information-determined buffers could equal flat buffer service level performance at reduced cost in some cases due to a reduction in over inventory.