laborious, which is why multivariate statistics and machine learning techniques have been successfully used to predict the behavior of variables and find hidden patterns in the data. Niquini and Costa (2020a, 2020b) and Niquini et al. (2023) developed several supervised regression models with this goal, while the latter also applied a recursive feature elimination technique to keep only the most relevant and reliable variables in the dataset. On the other hand, Costa et al. (2025) employed cluster analysis to determine reliable geometallurgical domains and then used hierarchical indicator kriging to model a mineral deposit, while Niquini et al. (2025) proposed an unsupervised approach to generate clusters from scarce geometallurgical data. All these examples demonstrate that, despite the challenges, it is possible to incorporate robust and reliable geometallurgical data into the mine planning pipeline, although the difficulties should not be overlooked. 4. CONCLUSIONS This study investigated the impact of incorporating geometallurgical variables into the mine planning framework. Four different scenarios were proposed: the first involved the traditional valuation of the blocks, while the others incorporated two geometallurgical variables in distinct combinations. The numerical experiments demonstrated that the simulated annealing metaheuristic can be a robust tool to achieve feasible solutions to the mine sequencing problem in an acceptable computational time, as the most time-consuming experiment required approximately 17 minutes. When comparing the different scenarios, it is conspicuous that ignoring the geometallurgical variables may result in simplistic and optimistic plans that will not achieve the expected return, given that the processing plant would not operate continuously with the same recovery and specific energy over the 16 years of life of mine (LOM). Future works should incorporate new constraints into the mathematical model to ensure that the generated mining plans are fully operational, including those related to minimum mining width and period controlled stripping ratio. This methodology, focused on incorporating geometallurgical variables into mine planning, can also benefit from the use of different algorithms in model resolution, including greedy randomized adaptive search procedure (GRASP) or population metaheuristics such as genetic algorithms (GA) or particle swarm optimization (PSO). Finally, it is also possible to carry out a stochastic optimization approach, considering the incorporation of geological, economic and/or operational uncertainties, so that the generated mining plans would be even more robust and risk-averse. ACKNOWLEDGEMENTS The authors would like to acknowledge the support of the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) institution for partially financing this research through a post-doctoral program scholarship nº 88887.014998/2024-00. REFERENCES Bienstock, D., and Zuckerberg, M. (2010). Solving LP relaxations of large-scale precedence constrained problems. In: Integer Programming and Combinatorial Optimization. IPCO 2010. Berlin, Heidelberg: Springer. hiips://doi.org/10.1007/978-3-642-13036-6_1. Boland, N., Fricke, C., and Froyland, G. (2007). A strengthened formulation for the open-pit mine production scheduling problem. Available from: <hiips://optimization-
RkJQdWJsaXNoZXIy MTM0Mzk2