Integrated Production Schedule Optimisation for Sublevel Stoping Mines

Tactical mine planning is the process of determining the sequence and schedule of activities that will best achieve the value generated through the strategic mine plan. These objectives are generally based on the maximisation of Net Present Value (NPV). The scheduling process over both the medium and short term is still largely a manual process as current mathematical optimisation models contain a number of limitations. This is particularly the case for underground mining methods such as sublevel stoping which generally involves selecting production from the next available highest cashflow stope. In most cases, the scheduling problem is too complex to guarantee that the optimal result will be achieved through manual scheduling. The solution to the medium term schedule will traditionally form the starting point from which to commence short term scheduling. Due to the segregation between the two phases only a local optima is possible. While maximising NPV is generally the primary objective of medium term scheduling, short term scheduling tends to focus on value improvements gained through the operational efficiencies resulting from consistently meeting production targets. This thesis describes and presents three mathematical optimisation models utilising mixed integer programming that have been developed for application to new and existing sublevel stoping operations. The locally optimal medium term production scheduling model optimises the production schedule over the medium term by maximising NPV. The locally optimal short term scheduling and machine allocation model generates optimal short term production schedules with all corresponding machine allocations for the purpose of minimising deviation to a target mill feed grade. These models are finally integrated to produce the globally optimal integrated production scheduling model, for the purpose of maximising final NPV while considering each activity across both scheduling horizons. A conceptual sublevel stoping operation comprising of 100 stopes is used for trialling three scheduling approaches incorporating the newly developed models. An initial ‘first pass’ manual scheduling approach achieves an NPV of $109,486,627. Segregated medium and short term schedule using the newly developed models achieve an NPV of $120,165,839. The final scheduling approach involved integrating the two scheduling horizons using the globally optimal integrated production scheduling model to achieve an NPV of $121,031,325.