Key Takeaways
- Long-term mine planning is one of the more genuinely AI-native categories on this list. Maptek Evolution runs many scheduling scenarios in parallel on cloud compute, and the solutions learn from each other to improve iteratively.
- Whittle (Enterprise Optimizer) and Deswik.Sched are the other major players, each with its own take on schedule optimization.
- Evolution differs from a traditional single-pass optimizer. Its approach is closer to evolutionary or genetic search than a one-shot LP solve.
- Marked “Yes” (not “Emerging”) because this is a shipped, production capability, not a research prototype.
TL;DR
For long-term mine planning with AI, run your schedules through a tool like Maptek Evolution. It evaluates many scheduling scenarios in parallel and improves them with evolutionary search, instead of producing one static LP schedule. Budget real setup time, because encoding your constraints and objectives properly is most of the work.
What Is Long-Term Mine Planning?
Long-term mine planning decides what gets mined, in what order and when, over the life of the mine. It usually covers years or decades. Planners start from the block model, which holds the grade and other attributes of every block. They then set the final pit limits, the pushbacks or stages, and a yearly production schedule.
The goal is usually to maximize net present value (NPV) while respecting real constraints. These include mining and processing capacity, cut-off grades, access sequencing and blending targets. Those targets later drive day-to-day ore blending optimization. Classic tools solve this with the Lerchs-Grossmann pit algorithm and linear or mixed-integer programming. However, the number of possible schedules is enormous, so any single solve explores only a small part of it.
How Do I Optimize Long-Term Mine Planning With AI?
The approach has three parts: understand the limits of single-pass solvers, search more widely, then pick a tool that fits your workflow.
Moving beyond one schedule at a time
Traditional tools solve for one schedule at a time. You set constraints, run a linear or mixed-integer optimizer and get one output. Then you tweak constraints and run again. It works, but it explores the solution space slowly. It can also get stuck at a local optimum that a human planner then has to second-guess.
Using evolutionary search for long-term mine planning
Maptek Evolution takes a different approach. It runs many candidate schedules simultaneously on hybrid cloud compute. True to its name, it uses an evolutionary process: better-performing schedules inform the next generation of candidates. As a result, it explores a much wider space of sequencing and scheduling decisions than a traditional solver in the same time. It is also less likely to hand you a single “optimal” answer that is actually a local optimum.
Choosing a long-term mine planning tool
Whittle’s Enterprise Optimizer and Deswik.Sched are the other established options, each strong in its own niche. Whittle is historically strong on pit-shell and cut-off grade optimization theory. Deswik is strong on integration with its broader scheduling and design suite. In practice, the key test is less the algorithm and more how well a tool fits your existing mining block model and mine design workflow.
What to Expect in Practice
Expect a real onboarding cost whichever tool you pick. These are enterprise planning tools, not plug-and-play scripts. Getting your constraints and objective function properly encoded takes real setup time.
A long-term plan is also only as good as its inputs. Therefore, block model quality, recovery assumptions and price forecasts matter as much as the optimizer. Plans are revisited regularly, so treat the schedule as a living model. Finally, compare actual output with the plan. Mine production forecasting closes that loop on the short-term side.
FAQs
How does AI improve long-term mine planning?
AI-driven tools such as Maptek Evolution explore many schedules in parallel instead of one at a time. Better schedules guide the next round of candidates. This widens the search and reduces the risk of settling on a local optimum.
What is the difference between Evolution and Whittle?
Maptek Evolution uses evolutionary search across many parallel scenarios. Whittle’s Enterprise Optimizer is historically strong on pit-shell and cut-off grade optimization theory. The right choice depends on your deposit, constraints and existing software stack.
What data does long-term mine planning need?
It needs a block model with grade and other key attributes for each block. It also needs costs, recoveries, prices, capacities and geotechnical slope limits. Blending and sequencing constraints complete the picture.
Is AI long-term mine planning ready for production use?
Yes. Maptek Evolution is a shipped, production capability rather than a research prototype. However, results still depend on careful setup of constraints and objectives.
Try It With Geocluster
Comparing evolutionary scheduling approaches like Evolution against traditional LP-based optimizers is a research-heavy decision. It depends on your specific deposit and constraints. The Geocluster Research Harness is built to help teams evaluate AI-driven planning tools like this rigorously before committing. You can also read how the Geocluster Research Harness works.