traditional economic function based on fixed cutoff, revenue, and costs; ii) in the second, there is variation in block recovery; iii) in the third, there is variation in processing cost as a function of specific energy; iv) in the fourth, finally, there is simultaneous variation in recovery and specific energy in blocks costs. The mine sequencing problem is then formulated as a direct block sequencing (DBS) model and solved by a simulated annealing metaheuristic for these four scenarios, and the results are compared in terms of economic return and robustness. Therefore, this study seeks to better understand the impact of incorporating geometallurgical variables into the mine planning framework and to popularise this approach to improve the practices currently employed in the mineral industry. 2. METHODOLOGY 2.1 Marvin dataset The numerical experiments in this study were developed on the Marvin dataset (Minelib, 2025), a notorious synthetic dataset of a polymetallic deposit (gold and copper) with 53,271 blocks measuring 30 x 30 x 30 meters. The original dataset contains information on block coordinates, tonnage, gold content (ppm), copper content (%), and the economic return from processing each block. In this study, an updated version of Marvin was used. This updated dataset has two geometallurgical variables, recovery (%) and specific energy (kWh/t) (Mazzinghy et al., 2025). 2.2 Modelling the block value in different scenarios 2.2.1 Scenario 1: traditional block value For this scenario, Equation (1) was implemented with the block value ($) calculated for gold and copper selling prices of = 12 $/ and = 20 $/ , respectively; gold and copper recoveries of = 0.6% and = 0.88%, respectively; gold and copper selling costs of = 0.2 $/ and = 7.2 $/ , respectively; a mining cost of = 0.9 $/ ; and a processing cost of = 4.0 $/ , as adopted by Minelib (2025). Thus, only the coordinates, tonnage, and grades of blocks vary in this scenario. Furthermore, the cutoff grade policy stipulated that a block would be destined for the waste dump if no metallic content was identified ( , = 0 and , = 0,∀ ∊ ), while blocks with any metallic content would be subject to the decision between being sent to the processing plant or to the waste dump. 2.2.2 Scenario 2: incorporating variable recovery into the block value For this scenario, Equation (1) was adapted for different recovery rates → , ∀ ∊ , ∀ ∊ . The values of , were obtained from Mazzinghy et al. (2025), who generated synthetic geometallurgical recovery models for three different lithologies. Figure 2 presents these recovery models (%) for Au (ppm) and Cu (%) grades.
RkJQdWJsaXNoZXIy MTM0Mzk2