Mine planning updates with revised information

In parallel to growth in software and operation research technologies in the last 20 years, mine design andproduction scheduling software became an important element of mineral industries. This is currently verydynamic, global and competitive industry. As a part of this industry, Flairbase also develops, markets andsupports various software products. However, mine design and planning software companies have experiencedwith new challenges over time. When a mining operation begins, inevitable inconsistencies between the plangenerated by the software and the reality are generally observed. One of the reasons for this is the uncertaintiesof the data used in the scheduling process. These uncertainties arise from two sources: (i) Sparse data used inspatial characterization of the deposit. Grade input is generated by an estimation or simulation method, whichis an uncertain image of the mineral deposit. (ii) Either long term projections or stochastic processes are used toobtain the financial scenarios (e.g. price, costs and discount rate) attributed to unknown future events. There arealso many assumptions made throughout deposit characterization and optimization. On such an occasion, thedecision maker needs a tool to interpret on the qualities of estimations, simulations, assumptions andparameters selected in the initial planning stage. If the software has a feature, which performs in an uncertainand complex environment in such a way as to respond to the inconsistencies, previous information can berevised using new information obtained by new samples from blasting drills, reverse circulation drilling andgeological observations. Flairbase expressed an interest to collaborate with McGill researchers to producedynamic and responsive planning practices such that a feedback on the feasibility of the existing plan isprovided, and if necessary this plan is updated. Thus, mine planning software company will have a distinctivefeature in the market. The research methodology is based on the incorporation of reinforcement learning intomine optimization. This ENGAGE project will help Flairbase in establishing long-term research collaborationwith the applicant's research team in which new planning practices will be introduced to the mining industry.