270 • States ({): In this context, a state j> ∈J at time M represents a complete description of the mining environment at a given moment. This includes, among other factors, the remaining resources, geological conditions, available equipment, and any other relevant parameters influencing mining operations. • Actions (|): Actions k> ∈}(j>) are operational decisions that the agent (decision-maker) can execute in state J>. These may include decisions such as the volume of material to extract, the type of technology to use, transportation routes, and other operational strategies like cut off police. Each action influences the transition from the current state to the next. • State Transition Function (~):The transition function L(j>?!,k>?!|j>,k>) defines the probability of transitioning from state (j>) to state (j>?!) after executing action (k>). This function captures the stochastic dynamics of the mining environment, incorporating uncertainties such as geological variations, machinery failures, and environmental conditions. • Rewards (): The reward function V(j>,k>) assigns a numerical value to the immediate reward obtained after executing action k> in state j> .In the mining context, this typically corresponds to the cash flow generated by mining operations, adjusted by operational costs and other economic factors. This formulation of the mining problem as an MDP enables the generalization of the approach to other problems that can be modeled under the premises of states, actions, and rewards. Essentially, any problem that can be described using these elements is amenable to being solved through Reinforcement Learning techniques. 3. CASE OF STUDY This section presents a detailed case study in the mining sector, focusing on the application of advanced data analysis techniques and machine learning algorithms to optimize mineral resource exploitation. The main objective is to maximize the Net Present Value (NPV) of the project, considering the operational and economic constraints of the mining environment. Using Reinforcement Learning approach, the goal is to develop production strategies that are not only economically viable but also technically feasible, enabling more efficient management of mineral resources. In Figure 4 it can be see the relationship between cut off grade, tonnage and average grade of resources. Also, in Figure 5 it can be see the economic parameters of mining operations that include processing, mining cost, prices of commodities and other parameters that permit evaluated in economic terms the feasibility of the project. Figure 4 - Geology Resources
RkJQdWJsaXNoZXIy MTM0Mzk2