Short- and medium-term optimization of underground mine planning using constraint programming

In the past few years, the mining industry has seen a lot of operational changes. Digitalization and automation of many processes have paved the way for an increase in its general productivity. In keeping with this trend, this article presents a novel approach for optimizing underground mine scheduling for the short- and medium-term. This problem is similar to the Resource-Constrained Project Scheduling Problem, with the difference that all task completions are optional. The model uses Constraint Programming principles to maximize the Net Present Value of a mining project. It plans work shifts for up to a year in advance, considering specialized equipment, rock support and operational constraints. This is the first published paper using optional variables to model optional tasks in a real-life application. Results from its applications to datasets based on a Canadian gold mine demonstrate its ability to find optimal solutions in a reasonable time. A comparison with an equivalent Mixed Integer Programing model proves that the Constraint Programming approach offers clear gains in terms of computability and readability of the constraints.