A HYBRID PARTICLE SWARM OPTIMIZATION SOLUTION TO CONSTRAINED LONG-TERM PRODUCTION SCHEDULING AT OPEN PIT MINING.

Optimization plays a key role in various fields of science specially engineering to simplify the decision making process. Long-term production scheduling (LTPS) of open pit mines is a large-scale optimization problem to determine the block extraction sequence to maximize the net present value considering technical, economic constraints. For many optimization problems, finding the best solution is intricate and timeconsuming. In such cases, a combination of mathematical methods and meta-heuristic techniques provides a good solution in a reasonable time. This paper presents a hybrid model between lagrangian relaxation (LR) and particle swarm algorithm (PSO) to solve the LTPS problem under the condition of grade uncertainty. We suggest to apply the LR method on the LTPS problem which to improve its performance speeding up the convergence and also, PSO is used to update the lagrangian multipliers. In this paper, firstly, we propose new diversification techniques for the second approach in order to get better results and secondly, we propose a new promising approach combining the two latter ones. The results of the case study demonstrate the LR approach in solving large-scale problem and produce an acceptable solution is more effective than traditional linearization method. In addition, the suggested hybrid strategy based on PSO, showed better performance than existing methods. The results obtained show that the extended version has brought substantial improvements compared with the first approach. In fact, the results are too close to the best results obtained in the literature. This is due essentially to the diversification added. [ABSTRACT FROM AUTHOR]