A NON-LINEAR OPTIMISATION APPROACH TO IN-PIT HAUL ROAD DESIGN

Abstract

Truck haul costs, as one of the predominant operational costs for mining and quarrying operations, are known to be heavily dependent on the design parameters of haul roads. Furthermore, in-pit haul road design parameters determine the pit limits and therefore, the potential feasibility of the mining operation. Thus, when in search of an optimal solution in terms of in-pit haul roads, one should primarily consider the location of the in-pit haul road, its design features and the deriving operational costs regarding extraction and haul costs. A suitable objective function in this case may be the undiscounted profit for the ultimate pit design. However, for each considered scenario, truck and excavator operational costs can be calculated using simulation techniques for better accuracy. Furthermore, finding an optimal solution requires the execution of a reliable and efficient algorithm, depending on the shape of the objective function. Hence, a non-linear optimisation approach was proposed in this paper for solving the in-pit haul road optimisation problem, based on a simulation of the materials allocation, which was used for calculating the objective function. Design parameters were assumed to be predetermined, while the only variable used for finding an optimal solution was the location of the in-pit haul road inside the pit contour. In addition, two 1-D algorithms were compared for finding the optimal solution (Search with accelerated step size and 1-D Simplex method). Furthermore, two regression models are proposed (Multiple Linear Regression – MLR and Non-Linear Regression - NLR), which could identify the more feasible region for the in-pit haul road location and reduce the number of iterations required for convergence.