Scholarlink

<p>This research addresses the design of haulage ramps within open-pit mines, a critical aspect of the mine design process, through a binary programming model that minimizes total construction and transport costs. Ramps are represented as interconnected segments across pit levels, ensuring compliance with gradient limits. Alongside this formulation, three heuristic algorithms were developed and evaluated: a greedy approach (MEGAP algorithm), a semi-randomized heuristic (MEGRASP), and a Local Search (LS) procedure. MEGRASP balances randomness and local refinement, while LS systematically improves MEGAP solutions by exploring all valid segment configurations. These methods were tested on two large-scale real-world datasets—McLaughlin and Copper Mine. Results show that LS consistently outperforms MEGAP, reducing ramp costs by up to 50% while maintaining geometric feasibility and computational efficiency. MEGRASP also proved competitive, delivering significant cost savings with minimal sensitivity to algorithmic randomness. A novel convergence criterion based on a ratio of parameters was introduced to guide MEGRASP configuration. Compared to manual or CAD-based design approaches, the proposed methods offer greater accuracy, scalability, and automation. These results underscore the value of optimization-based tools for practical mine planning and open avenues for further application in underground mining, construction, and other infrastructure design contexts.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Using a Semi-randomized Heuristic and Local Search Methods to Solve the Open-pit Haulage Ramp Design Problem

  • 期刊论文

  • 发表日期：

    2026年2月19日

查看全文
  • Juan L. Yarmuch,
  • Erick Sanhueza,
  • Joaquín Córdova,
  • Diego Mancilla,
  • Rodolfo Gaínza,
  • Kimie Suzuki

摘要

This research addresses the design of haulage ramps within open-pit mines, a critical aspect of the mine design process, through a binary programming model that minimizes total construction and transport costs. Ramps are represented as interconnected segments across pit levels, ensuring compliance with gradient limits. Alongside this formulation, three heuristic algorithms were developed and evaluated: a greedy approach (MEGAP algorithm), a semi-randomized heuristic (MEGRASP), and a Local Search (LS) procedure. MEGRASP balances randomness and local refinement, while LS systematically improves MEGAP solutions by exploring all valid segment configurations. These methods were tested on two large-scale real-world datasets—McLaughlin and Copper Mine. Results show that LS consistently outperforms MEGAP, reducing ramp costs by up to 50% while maintaining geometric feasibility and computational efficiency. MEGRASP also proved competitive, delivering significant cost savings with minimal sensitivity to algorithmic randomness. A novel convergence criterion based on a ratio of parameters was introduced to guide MEGRASP configuration. Compared to manual or CAD-based design approaches, the proposed methods offer greater accuracy, scalability, and automation. These results underscore the value of optimization-based tools for practical mine planning and open avenues for further application in underground mining, construction, and other infrastructure design contexts.