Use of Machine Learning Algorithm Models to Optimize the Fleet Management System in Opencast Mines

In surface mining operations, the dumper haulage system contributes the most in total operating cost of any mine. It is estimated that an average mining company spends around 50% to 60% in this truck haulage system only. So utmost priority should be given to keep up an effective haulage framework. So, to reduce the cost of operation the dumpers must be allocated and dispatched efficiently. The haulage systems should be designed in such a manner that the availability, performance and utilization of the dumper and shovel are maximized, which ultimately yield in high production and reduction of operating cost. So, in this paper to enhance the productivity of truck haulage system an attempt is made to minimize the cycle time of dumpers and allocate an optimized number of dumpers to one shovel so that the idle time of dumpers can be minimized. In determining the cycle, time of dumpers predicting the travelling time in different situation is given utmost importance. For the machine learning models are used which help in predicting the travelling time in different atmospheric situation of the mine. This approach of integrating the machine learning methods in minimizing the cycle time will provide a proper estimation of performance measure, truck scheduling and finally an optimized truck dispatch system.