Stochastic open pit optimization – gold project in Canada

Client & context

The project involved a gold project in Canada, owned by McEwen Mining, undertaken during 2018–2019.

The client wanted to understand how orebody geological uncertainty could impact the economic value of an open-pit mining project—beyond what traditional single-model pit optimisation and scheduling could show.

Challenge

Traditional pit optimisation assumes a single deterministic geological model. In reality, however:

• Grade and geometry vary at multiple scales
• Local misclassification can impact both short-term production and long-term value
• A pit that looks optimal on paper can perform very differently in practice

The client needed to:

• Incorporate geological uncertainty (multiple realisations of the orebody) into both pit design and life-of-mine (LOM) scheduling
• Understand how different orebody realisations affect NPV, production profiles and risk
• Develop feasible, implementable production sequences, not just theoretical shells

In short: they needed a way to design and schedule the pit that captures maximum upside while controlling geological risk.

Our approach

1. Simulating geological uncertainty
   Gosselin Mining worked with the client’s geological data to generate stochastically simulated orebody models using sGems. The focus was on:
   • Creating multiple, geostatistically sound realisations of grade and geometry
   • Validating the simulations to ensure they were realistic and decision-worthy
   • Checking that each model could be carried through to optimisation without technical issues
2. Stochastic open-pit optimisation and LOM scheduling
   Once the simulated orebody models were validated, they were imported into a stochastic optimisation platform to perform:
   • Stochastic pit optimisation that considers multiple geological scenarios
   • Life-of-mine production scheduling based on those scenarios
   • Evaluation of risk profiles for different designs and sequences

Significant effort went into guiding the optimiser to produce practical LOM sequences:

1. • Analysing risk profiles associated with each simulated orebody
   • Visualising schedules and mine plans across scenarios
   • Identifying and resolving potential bottlenecks early (e.g. capacity, blending, cut-offs)
2. Risk-informed planning and visualisation
   Through detailed visualisation and scenario comparison, the client could see:
   • Where geological risk was most concentrated in the pit and schedule
   • How different sequences and cut-off strategies behaved under varying geological outcomes
   • Which plans offered the best trade-off between value and risk over the life of the mine

Results & value for the client

The stochastic optimisation study showed clear advantages over traditional deterministic planning:

• A higher-value life-of-mine plan compared with conventional scheduling methods used as a benchmark
• Better production planning, with sequences that are robust across multiple geological realisations
• Reduced geological risk over the life of mine, thanks to a design and schedule that explicitly account for uncertainty
• Improved decision-making, as management could see both upside potential and downside exposure in a quantified way

The project was successfully completed within the agreed schedule, using advanced mining software to integrate simulation, optimisation and scheduling in one workflow.

For more technical details, the study is documented in the Swedish Governmental Agency for Innovation Systems (Vinnova) project database under “Stochasctic Mine Design”.

Reference

1. GERAd (Onine) A platform for optimizing mining complexes with uncertainty. Available at: https://www.gerad.ca/fr/papers/G-2016-96 (Accessed on 18 July 2025)