Presentation Information
[3111]An Optimization-Based AI Agent for Short-Term Mine Planning in Gold Heap Leach Operations
○Angelo Giovanni Paravecino Tena1 (1. Hokkaido University)
Chairperson: 才ノ木敦士（熊本大学）
Keywords:
Short-term mine planning,Artificial intelligence,Optimization
Short-term mine planning in open-pit gold operations requires the continuous evaluation of multiple mining fronts under changing operational conditions. Mine planners must balance production targets, gold grade requirements, equipment availability, and spatial constraints while making timely decisions. As the number of feasible alternatives increases, the planning process becomes increasingly complex and dependent on individual expertise.
This study presents an optimization-based artificial intelligence agent designed to support decision-making in short-term mine planning for a gold heap leach operation. The proposed agent integrates information from block models, current topography, mine designs, equipment availability, and productivity parameters to identify and prioritize mining fronts capable of meeting weekly production objectives. A mathematical optimization framework is used to evaluate alternative mining scenarios and recommend extraction sequences that maximize compliance with tonnage and grade targets while considering operational constraints.
The methodology was implemented using data from an operating open-pit gold mine. Available mining fronts were characterized according to their tonnage, average gold grade, accessibility, and equipment requirements. The agent generated and ranked alternative plans based on their expected operational performance and alignment with production goals. The recommended solutions were compared against conventional planning practices using key performance indicators related to tonnage achievement, grade control, and equipment utilization.
Results demonstrate that the proposed approach can rapidly evaluate multiple planning alternatives and provide consistent recommendations for front selection under dynamic operating conditions. The framework enhances transparency in the planning process and supports planners in identifying opportunities to improve production compliance. The proposed AI agent represents a practical decision-support tool that complements engineering judgment and contributes to more efficient short-term mine planning in heap leach operations.
This study presents an optimization-based artificial intelligence agent designed to support decision-making in short-term mine planning for a gold heap leach operation. The proposed agent integrates information from block models, current topography, mine designs, equipment availability, and productivity parameters to identify and prioritize mining fronts capable of meeting weekly production objectives. A mathematical optimization framework is used to evaluate alternative mining scenarios and recommend extraction sequences that maximize compliance with tonnage and grade targets while considering operational constraints.
The methodology was implemented using data from an operating open-pit gold mine. Available mining fronts were characterized according to their tonnage, average gold grade, accessibility, and equipment requirements. The agent generated and ranked alternative plans based on their expected operational performance and alignment with production goals. The recommended solutions were compared against conventional planning practices using key performance indicators related to tonnage achievement, grade control, and equipment utilization.
Results demonstrate that the proposed approach can rapidly evaluate multiple planning alternatives and provide consistent recommendations for front selection under dynamic operating conditions. The framework enhances transparency in the planning process and supports planners in identifying opportunities to improve production compliance. The proposed AI agent represents a practical decision-support tool that complements engineering judgment and contributes to more efficient short-term mine planning in heap leach operations.