Search

With globalisation, the need to have better coordination of production, inventory, and distribution decisions has become increasingly important for industrial gas companies in order to remain competitive in the marketplace. The consolidation of such decisions has an immense impact on the performance of their supply chains and overall profitability. In this thesis, mixed-integer linear programming (MILP) models are proposed to address integrated planning for industrial gas supply chains. The MILP models consider complicating features in the supply chains, e.g., production, inventory management under vendor managed inventory (VMI) system, supply contracts, and transportation schedule. Since solving the integrated problems with this high level of detail requires expensive computational costs when dealing with large-size problems, a hierarchical solution strategy is developed to obtain optimal solutions within reasonable computational times. Next, allocation of transportation resources (i.e., railcars and trucks) is also considered to improve the transportation efficiency and total operating cost simultaneously. Optimisation approaches are proposed based on mixed integer linear fractional programming (MILFP) and multi-objective optimisation (MOO) models. As a solution procedure to the MILFP model, Dinkelbach's algorithm and reformulation-linearisation method are adopted, whereas the ε-constraint method is used for the MOO model. The optimisation problem for the production and inventory routing planning during the COVID-19 pandemic is also addressed. Because the vehicle routing optimisation problem is NP-hard, tailored solution methods are developed based on column generation principles or metaheuristics to tackle larger problem instances. For each of these cases, the applicability of the developed models and solution techniques are tested using real-world case studies of CO2 supply chain planning in the US and O2 supply chain planning in the UK. Furthermore, the capability of models capturing all features mentioned is discussed with the effectiveness of the solution methods in the case studies.