Last couple of years, we have during several R&D projects (e.g. the SIMS project) integrated the SimMine discrete-event mine simulator with our production scheduling system. It has been extended beyond a long-term mine simulator by including more logic related to detailed simulation of mining activities. By integrating with our scheduling system, the production in this “virtual mine” can now be coordinated by short-term schedules that are continuously revised using our algorithms (close to lvl 6 fully automated SIC as stated in the guideline by GMG).
In the early 2010’s, we showed that it’s possible to produce vastly more efficient short-term production schedules using algorithms rather than manual labor. However, while optimizing is one thing, the real struggle in automatic rescheduling of real-life production is to robustly adhere to soft constraints, corner cases, and personal preferences. All within a couple of seconds to make sure that the operator will indeed use the proposed system as decision-support.
A paper based on these projects is in submission for publication. Hopefully it will available later this year.
Max Åstrand 