一种基于强化学习的矿山开采方案多目标优化方法
By constructing a mine geological model and using reinforcement learning trial and error, the problems of insufficient dynamic coupling modeling and inability to adjust safety constraints in multi-objective optimization of mines were solved, realizing efficient adaptive mine mining decision support and improving the accuracy and safety of mining schemes.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHANGCHUN GOLD DESIGN INST
- Filing Date
- 2026-02-05
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies lack modeling of the dynamic coupling relationship between geological resources, engineering feasibility, and safety risks in multi-objective optimization of mines. This makes it difficult for optimization models to accurately reflect the state evolution of the real mining environment. At the same time, traditional reinforcement learning frameworks cannot achieve adaptive adjustment of safety constraints, which can easily lead to policy convergence bias or violation of safety regulations.
A mine geological model is constructed, and the ore body resource layer, mining engineering constraint layer, and safety risk field layer are arranged in a layered database to generate a mine environmental state database. Mining schemes are generated through reinforcement learning trial and error, and adaptive adjustment of safety constraints is achieved by combining Lagrange multiplier tiered updates.
It enhances the authenticity of decision-making and the reliability of strategies, strengthens the balance between multi-objective optimization and safe production, and realizes dynamic adjustment of safety constraints and efficient adaptive decision support.
Smart Images
Figure CN122046977B_ABST