ABSTRACT

In mines planning, the long-term production scheduling problem (LTPSP) in open-pit mines is considered as a significant issue. It also specifies the distribution of cash flow during the course of the mine-life. Actually, LTPSP is a large-scale optimisation problem including large data-sets, multiple constraints, and uncertainty in the input factors that, has to be solved in a reasonable time. LTPSP, despite the valuable efforts of researchers, has not yet been well resolved. In this paper, hybrid models have been offered by the Lagrangian relaxation (LR) method with meta-heuristic methods, bat algorithm and particle swarm optimisation for solving the LTPSP due to the deterministic assumption and concerning the grade uncertainty. To bring update the Lagrange multipliers, the meta-heuristic algorithms have been applied. In terms of cumulative net present value, average ore grade, and computational time in a 12-year production period, the consequences achieved from the case studies point out that a solution close to optimisation can be presented by the LR-bat algorithm hybrid strategy in comparison with other methods. The results analysis has shown that the proposed method produces a near-optimal solution with a rational time that can be a good suggestion for utilising in the mining industry.

摘要

在矿山规划中，露天矿的长期生产调度问题（LTPSP）被视为一个重要问题。它还规定了矿山开采期间现金流量的分配。实际上，LTPSP是一个大规模的优化问题，其中包括大数据集，多重约束以及必须在合理时间内解决的输入因素的不确定性。尽管研究人员做出了巨大的努力，但LTPSP尚未得到很好的解决。在本文中，由于具有确定性假设并且涉及坡度不确定性，通过拉格朗日松弛（LR）方法，元启发式方法，蝙蝠算法和粒子群优化提供了混合模型来求解LTPSP。为了更新拉格朗日乘数，已应用了元启发式算法。在12年生产期的累积净现值，平均矿石品位和计算时间方面，案例研究得出的结果指出，LR-bat算法的混合策略可以在以下情况下提供接近优化的解决方案：与其他方法的比较。结果分析表明，所提出的方法能够在合理的时间内产生接近最优的解决方案，这对于在采矿业中的应用是一个很好的建议。