Mining and optimisation
AI and Autonomy Lab project
Project description
Mine planning is complex with short, medium and long-term objectives that are often competing rather than complementary. Increasing capability in data collection and management, moreover, is enabling faster and more frequent updating of orebody models. Fast, and predominantly automated, scheduling tools can take advantage of these frequent model updates by developing new, or adjusting existing plans, as required. This contrasts with current practice in which the time consuming nature of planning prevents this kind of agility.
The Mining & Optimisation team within the AI & Autonomy Group has the capabilities to tackle these complex optimisation problems, with expertise in mathematical programming, discrete optimisation, classical planning, and stochastic optimisation. We have a track record of developing novel algorithms, and applying decomposition, aggregation and disaggregation, to complex scheduling problems.
In a prior collaboration with Rio Tinto Iron Ore (RTIO), we have developed efficient algorithms for generating multiple realistic short-term production plans for open-pit iron ore mines, optimised with respect to a series of prioritisable objectives. A large number of distinct plans can be generated in minutes, each satisfying all relevant constraints on equipment and plant utilisation, and KPIs on production. Our collaboration with RTIO has additionally resulted in algorithms for short-term scheduling across a network of mines, each contributing to a set of blended products, and hybrid optimisation methods for planning at daily and weekly timescales. We have experience in the mathematical modelling of key planning problems across all horizons, from short- to long-term.
Our researchers have collaborated with the Hunter Valley Coal Chain to develop an optimization-based logistics planning system integrating train scheduling, stockpile management, and vessel scheduling. This system-wide, coordinated approach deviates from earlier efforts that have focused on individual components of such supply chains, and integrates constraint programming, mixed integer programming, and local search methods.
Our research agenda and interests include:
- Multi-horizon planning
- Linking real-time control and data collection with planning tools
- Modelling and reasoning with uncertainty.
Multi-horizon planning
Current practice, and software tools, do not provide adequate support to mine planners for analysing the impact of decisions at one planning horizon on others. A new generation of planning technology is required to realise effective multi-horizon planning.
Real-time control, data collection, and planning
Fast and automated scheduling tools provide a means of adapting plans in response to new data as it arrives. This leads to interesting research questions, however, such as when and how to adapt plans in response to variation in predictions and their uncertainty.
Modelling and reasoning with uncertainty
We are interested in developing algorithms that are able to find a series of robust plans within reasonable time frames – drawing upon the fields of robust optimisation, simulation optimisation, and contingency analysis.
Researchers in the Mining & Optimisation team within the AI & Autonomy Group form part of a larger group of academics across the Faculty of Engineering and Information Technology working on mining-related problems. The Faculty of Engineering and Information Technology’s Sustainable Resources platform works with explorers, miners, and mining equipment, technology and service providers to tackle the sectors most challenging problems.
Researchers
Publications
Blom, M.; Pearce, AR.; Stuckey, PJ. (2018) “Multi-objective short-term production scheduling for open-pit mines: a hierarchical decomposition-based algorithm” Engineering Optimization pp: 1-18. DOI:10.1080/0305215X.2018.1429601
Blom, M.; Pearce, AR.; Stuckey, PJ. (2018) “Short-term planning for open pit mines: a review”International Journal of Mining, Reclamation and Environment pp: 1-22. DOI:10.1080/17480930.2018.1448248
Blom, M.; Pearce, AR.; Stuckey, PJ. (2017) “Short-term scheduling of an open-pit mine with multiple objectives” Engineering Optimization TAYLOR & FRANCIS LTD. pp: 777-795. DOI:10.1080/0305215X.2016.1218002
Blom, ML.; Pearce, AR.; Stuckey, PJ. (2016) “A Decomposition-Based Algorithm for the Scheduling of Open-Pit Networks Over Multiple Time Periods” Management Science INFORMS. pp: 3059-3084. DOI:10.1287/mnsc.2015.2284
Harabor, D.; Stuckey, PJ. (2016) “Rail Capacity Modelling with Constraint Programming” Quimper, CG. (Ed.) 13th International Conference on Integration of Artificial Intelligence and Operations Research Techniques in Constraint Programming (CPAIOR) Switzerland. SPRINGER INT PUBLISHING AG. pp: 170-186. DOI:10.1007/978-3-319-33954-2_13
Burt, CN.; Lipovetzky, N.; Pearce, AR.; Stuckey, PJ. (2015) “Scheduling with Fixed Maintenance, Shared Resources and Nonlinear Feedrate Constraints: A Mine Planning Case Study” Michel, L. (Ed.) 12th International Conference on Integration of Artificial Intelligence (AI) and Operations Research (OR) Techniques in Constraint Programming (CPAIOR) SPRINGER-VERLAG BERLIN. pp: 91-107. DOI:10.1007/978-3-319-18008-3_7
Blom, ML.; Burt, CN.; Pearce, A.; Stuckey, PJ. (2014) “A Decomposition-Based Heuristic for Collaborative Scheduling in a Network of Open-Pit Mines” INFORMS Journal on ComputingINFORMS. pp: 658-689. DOI:10.1287/ijoc.2013.0590
Belov, G.; Boland, N.; Savelsbergh, MWP.; Stuckey, PJ. (2014) “Local search for a cargo assembly planning problem” 11th International Conference on the Integration of Artificial Intelligence and Operations Research Techniques in Constraint Programming (CPAIOR) CH. Springer International Publishing. pp: 159-175. DOI:10.1007/978-3-319-07046-9_12
Lipovetzky, N.; Burt, CN.; Pearce, A.; Stuckey, PJ. (2014) “Planning for Mining Operations with Time and Resource Constraints” Steve Chien, Minh Do, Alan Fern, And Wheeler Ruml, . (Ed.) International Conference on Automated Planning and Scheduling (ICAPS) US. AAAI Press. pp: 404-412.