FREIGHT COST OPTIMISATION FOR A GLOBAL COMPANY

Abstract

This thesis focuses on the application of optimisation techniques to a real world large scale problem faced by a global mining company, XYZ. The problem which XYZ has been facing is that she has no proper optimisation procedure to minimise the monthly total freight costs of transporting aluminium ingots from its n smelters to m destination ports worldwide. Currently, she relies on employee experiences and judgemental heuristics to determine the weight of aluminium ingots to be shipped from each smelter to each port, and also the specific container liners to hire for each route. In addition, k of her smelters hire breakbulks, which are privately chartered vessels, to transport the aluminium ingots to a few dedicated ports. Some breakbulks require the fulfillment of a minimum capacity; otherwise penalty charges will be levied on the unfilled capacity. However, XYZ’s heuristic method leads to suboptimal decisions, where the final total freight cost incurred is more than the true optimised cost. In this thesis, an analysis of XYZ’s heuristic method and its shortcomings was made. Following, an optimisation model (transportation model) was designed in which for each route, the optimal weight of aluminium ingots to be shipped in order to minimise total freight costs was determined. Next, historical data was used to run the optimisation model in order to investigate the amount of cost savings with the implementation of the optimisation model over the heuristic method. The thesis then continued with extensions of the original problem, such as considering a number of commodities, the fulfillment of a minimum capacity for all breakbulks, the different prices charged by the various shipping liners for each route and the capacity constraint of each shipping line. Following, a generalised transportation model, which increases the flexibility for modelling XYZ’s future transportation problems, was also formulated. Finally, the thesis concluded with the importance of the optimisation model for XYZ and the future benefits that it may bring to her and the industry in general.