Models for a coal blending and distribution problem

This research is concerned with the study of typical coal flow operations in the coal industry, with a concentration on the blending and distribution problems. The problem here is to determine decisions and schedules for coal shipment to silos, cleaning and blending operations at the silos, and the subsequent shipment of coal to customers over a multi-period time horizon, so as to satisfy the demand at a minimum total operational cost. Three linear programming models of varying complexity and capability are developed to meet this objective. Based on a comparison of their structural properties and their accuracy versus computational requirements, recommendations are made for implementation. The final product envisioned is a computer-based decision tool which will serve as a mechanism for implementing cost-effective decisions in light of complex variations in production, ore quality and demand requirements over time. The models can be used to plan future growth and capital investment decisions, as well as to study the effect of various policies, by testing the sensitivity, feasibility and the cost of system operations under different perturbations of system configuration, data and demand specifications. Real operational data from the Westmoreland Coal Company is used for testing purposes.