Robust optimisation and modelling for rotables in supply chains for airline technical services : opportunities for improving decision-making

• Main Author: Möller, Eðvald
• Format: Dissertation
• Language: English
• Published: Imperial College London 2018
• DOI: 10.25560/81264

Supply chain management has been a subject of growing interest, both in terms of the related professional literature and in industrial practice; it has simultaneously undergone significant modification and expansion during the last decades. Companies have been working to increase their capabilities in a highly competitive market characterised by competition, which has been caused by the globalisation of industry, shortened product life cycles, increased flexibility and responsiveness to customer demand. These causes have been helped by advances in information technology and business optimisation. This thesis presents a methodology for supply management problems, focusing on achieving an adequate balance between detail and the complexity of the model, on the one hand, and usability on the other. The model approach that is presented focuses on predicting future demand and optimising procurement with minimum stock-out and inventory cost. We apply our approach to the problem of demand for spare parts in the airline industry, with regard to how to best schedule purchases and hold the inventory level minimum to meet the demand. We show how we can formulate a stochastic programming problem (based on a MILP). In order to increase the robustness of the plans, we use an iterative procedure which generates a number of alternative demand samples and uses them to evaluate the robustness. The iterative procedure is a computationally efficient alternative to stochastic programming approaches and it has been proven to increase the robustness of the forecasting plan for the problem under consideration. The data generator modules create demand scenarios derived from the stock-out observed between historical demand data and their corresponding forecasts. The tools established in this work use a modelling and simulation working side by side to attempt to reflect the best of both worlds. This was done to test the results given by the modeller against a more complex system to determine whether the performance of the system is still acceptable, thus allowing us to propose optimisation models to provide support for the relevant decisions. One of the models created can help to create a more sophisticated model in a supply chain and can be used as a more sophisticated tool in a wider supply chain simulation.