Geological variation may be managed by alternating between modes of operation. These modes provide an integrated response to changes in feed mineralogy, and other operational conditions within the mineral value chain. The decision to alternate between modes depends on current and forecasted stockpile levels. Moreover, the optimization of stockpile thresholds that would trigger a mode change is related to the classical RQ problem from inventory theory. Particularly for concentrators that are designed for blended feeds, there may be production intervals in which a particular ore class is at risk of stockout. Indeed, geological uncertainty causes stockout risk, which should be balanced with other key performance indicators, including throughput and recovery, and is a context for multiobjective optimization. Computations are presented in which a concentrator that had originally been designed for a certain ore blend, after years of successful operation, now requires alternate modes to realign its operations with geological forecasts.