Capturing the Value of Lithological Domain Uncertainty in Geometallurgical Mine–Plant Planning

Integrated mine–plant optimization should account for multiple geological uncertainties, yet practice typically considers only grade. Lithological domain uncertainty is a distinct source of geometallurgical variability because rock type governs comminution, recovery, cost, and blending. This paper extends a two-stage stochastic mine–plant planning framework to represent this uncertainty in a copper porphyry case study based on geological data from the Río Blanco–Los Bronces deposit in the Central Chilean Andes. Copper grade is fixed by block, while Truncated Gaussian Simulation generates realizations of tourmaline breccia and surrounding lithologies. The first stage optimizes a single long-term extraction schedule; the second evaluates recovery, processing cost at fixed throughput, Bond Work Index, and feed-composition compatibility under two adaptive operational modes. The framework combines parallelized Variable Neighborhood Descent with an embedded matheuristic for mass balance, plant capacity, and blending. Relative to a block-wise modal-domain deterministic benchmark over the same 100 validation scenarios, the stochastic plan increases mean net present value by US$2.38 million (3.28%), yielding an empirical value of the stochastic solution, and raises P10, P50, and P90 by US$2.27, US$2.28, and US$2.53 million, respectively. These results show that lithological domain uncertainty has measurable economic value in mine–plant planning even when copper grades are fixed.