An integrated production and direct shipment problem in a mining industry

This research aims to solve an integrated production, storage and distribution planning problem in a multi-level multi-product logistics system. The work was conducted in collaboration with a leading mining industry. The present research presents a mixed-integer linear program inspired by the classical Multi-Level Lot-Sizing model to solve the problem. It combines several aspects related to (1) multi-level and multi-product production units, (2) multi-level and multi-product storage units and (3) train transportation with time windows. Besides the standard features of lot-sizing problems, this work encompasses further aspects related to mining activity precisely and, more generally, to the structural specifications of continuous production processes. The objective is to hierarchically satisfy deterministic demand, minimise the total number of switch-offs of production units, and minimise total operational costs. On a sizeable real data-set obtained from our industrial partner, we demonstrate the benefit of integrating decisions in terms of improving demand satisfaction rates, increasing resource efficiency, and minimising total operational costs. In addition, we study the cost structure and provide a detailed cost sensitivity analysis. Finally, this work evaluates the upstream and downstream propagation of decisions in the logistics system and underlines the benefits in terms of flexibility offered by the global approach.