The Quest for Mining's Perfect Planner: How Scientists Found the Best Computer Program for Open-Pit Operations

Imagine standing at the edge of a massive mountain, knowing that somewhere within its rocky embrace lie precious metals worth millions of dollars. But here’s the catch: you can’t just blast away randomly. Every shovel of earth you move costs money, every piece of machinery has limits, and there are strict rules about which parts you can dig before others. Mining companies desperately needed the perfect digital planner to solve this three-dimensional puzzle. This is the story of how scientists embarked on a quest to find it.

The Ultimate Mountain Puzzle

Picture yourself trying to organize the world’s most complex three-dimensional jigsaw puzzle. You’re standing before a mountain that’s been divided into thousands of invisible blocks, like a giant Minecraft world made real. Each block contains different amounts of valuable ore, costs different amounts to extract, and some blocks can only be reached after removing others above them.

This isn’t just about digging holes in the ground. You’re planning years ahead, deciding which blocks to extract in which order to maximize profits while respecting all the rules of physics, safety, and business. It’s like being asked to plan the perfect road trip through a thousand cities, except the roads keep changing and each decision affects every future choice.

But What If Computers Could Master This Art?

Scientists set out on a mission: to find the ultimate computer program that could master this incredibly complex puzzle. Not just any computers, but specialized digital problem-solvers called “constraint programming solvers” - think of them as super-smart computer programs that excel at solving puzzles with millions of rules and restrictions.

The challenge they faced was enormous. Open-pit mining - the type where you create a massive crater to reach underground ore - involves what experts call the “Open-Pit Long-Term Production Planning Problem.” It’s like playing chess, but the board is three-dimensional, there are thousands of pieces, and every move must follow dozens of different rules simultaneously.

The Seven Digital Strategists

In their quest for the perfect mining planner, the research team gathered seven different computer programs, each representing a different approach to solving complex puzzles:

Gecode - The speed demon, built like a race car for constraint problems, specially tuned with efficient problem-solving techniques.

MiniZinc - The versatile translator, more like a universal language that can communicate with different solver engines underneath.

Mzn-g12cpx - The hybrid genius, combining two different types of problem-solving approaches like having both a mathematician and a logician work together.

Flatzinc, Mzn-g12fd, and Mzn-g12fdlp - The specialists, each designed for particular types of mathematical challenges.

Choco - The well-established veteran, built with state-of-the-art constraint-solving technology.

Here’s How They Ran The Test

The scientists created a digital test environment for their mining challenge. Imagine taking a mountain and slicing it into a three-dimensional grid of blocks, like cutting a giant cake into thousands of cubic pieces. Each block had specific properties:

• Value content: How much precious metal it contains
• Extraction cost: How much it costs to dig out
• Processing requirements: Whether the material needs special treatment
• Location constraints: Which other blocks must be removed first

Then they added real-world limitations, just like in actual mining operations:

• Processing plant capacity: You can only refine so much ore per day
• Mining equipment limits: Your excavators can only move so much material
• Grade blending rules: The mixture of different ore qualities must stay within certain ranges
• Safety slope requirements: You can’t create unstable cliff faces

The Quest Results: Champions Emerge

When the scientists tested all seven programs with 36 different mining scenarios of varying complexity, their quest finally revealed the champions. Gecode, MiniZinc, and Mzn-g12cpx emerged as the superior mining planners.

Gecode excelled because it was specifically designed for these types of constraint problems, with highly efficient “propagators” - think of them as specialized tools that quickly eliminate impossible solutions.

MiniZinc succeeded by being incredibly versatile, able to adapt its solving approach based on the specific problem characteristics.

Mzn-g12cpx impressed everyone by combining constraint programming with boolean logic solving, creating a hybrid approach that could find solutions faster by learning from previous attempts and eliminating dead ends more efficiently.

The performance differences were dramatic. While some solvers took over 10 hours to solve complex scenarios, the top performers could crack the same puzzles in minutes or even seconds.

This Means That…

This quest to find the perfect planner revealed crucial insights for the mining industry. Mining companies can now use these digital tools to make better decisions about:

• Which parts of their mining sites to develop first
• How to schedule their operations to maximize profits
• How to balance different types of ore to meet processing requirements
• How to plan mining operations years in advance with mathematical precision

Instead of relying solely on human experience and intuition, mining engineers now have access to computer programs that can explore millions of possible mining sequences and find the mathematically optimal solution.

In the Future…

These constraint programming techniques could revolutionize how we approach other complex resource extraction problems. Imagine applying similar digital problem-solvers to:

• Planning the layout of renewable energy installations
• Optimizing water resource management across regions
• Coordinating complex manufacturing processes
• Managing urban development projects with multiple constraints

The Bigger Picture

This research represents something profound about how we’re learning to solve humanity’s complex logistical challenges. We’re not just building faster computers; we’re creating digital specialists that can master the art of optimization in ways that would take human experts lifetimes to achieve.

Every time we extract resources from the Earth, we’re making decisions that affect the environment, the economy, and future generations. Having computer programs that can find the most efficient, least wasteful solutions means we can minimize environmental impact while maximizing the value we get from our natural resources.

The marriage of human wisdom and computational power in this research reveals a deeper truth about modern problem-solving. We’re no longer limited by the speed of human calculation or the boundaries of human intuition. Instead, we’re entering an era where the most complex puzzles our world presents - from mining mountains to managing cities - can be solved through the collaboration between human insight and digital precision. These seven computer programs didn’t just solve mining problems; they demonstrated how artificial intelligence can become our partner in making better decisions about our shared planet.

---

The Science Behind This Story

Published in: Ricardo Soto, Broderick Crawford, Boris Almonacid, Franklin Johnson, Eduardo Olguín (2014). Solving Open-Pit Long-Term Production Planning Problems with Constraint Programming - A Performance Evaluation. ICSOFT Conference. DOI: 10.5220/0005093900700077

What the scientists found:

• In their search for the perfect mining planner, they tested seven different computer programs on 36 mining scenarios
• Three programs (Gecode, MiniZinc, Mzn-g12cpx) emerged as the clear winners
• The champion programs could solve complex mining puzzles in minutes while others took hours
• Different programming approaches showed dramatically different performance levels, revealing which strategies work best for mining optimization

Why this research is important: This quest represented the first comprehensive search for the best constraint programming techniques for open-pit mining problems. Mining operations involve millions of dollars in decisions about which parts of the earth to excavate and when, so finding the ultimate planning tools can help companies make optimal decisions.

Who did this work: An international research team from Chilean universities (Pontificia Universidad Católica de Valparaíso, Universidad Autónoma de Chile, Universidad Finis Terrae, Universidad de Playa Ancha, Universidad San Sebastián) specializing in computational optimization and constraint programming applications.