When: Thursday 11th of November, 1pm AEDT
Where: This seminar will be presented online via Zoom, RSVP here.
Speaker: Haonan Zhou
Title: Adaptive Cross-entropy Method for Open-pit mine planning
Mineral deposits are discretised into standardised cuboids in open-pit mine operations, also known as “blocks”. Those blocks undergo extracting, processing, and stockpiling to form blended products of specified grades. Across a given time horizon, a production schedule decides which blocks are extracted, the extraction order, and the corresponding destinations, e.g., waste dumps, stockpiles, processing facilities, etc. The mining operations are constrained by a few factors such as equipment and or facility capacity, mining precedence and blending requirement. Due to the large number of blocks involved in a real-world mine site and the mixing at any intermediate junctions, the optimisation problem of open-pit mine production scheduling is complex by its nature and hence difficult to solve. The cross-entropy method is a two-phase iterative algorithm suitable for solving complex combinatorial optimisation problems. In this work, a novel adaptive cross-entropy algorithm is proposed to address a short-term production scheduling problem in an open-pit mine. More specifically, the cross-entropy method is utilised to make a probabilistic selection of blocks for complexity reduction.
Haonan Zhou is a PhD candidate in the Rio Tinto Centre for Mine Automation (RTCMA) at ACFR. His PhD is focused on integrated optimisation of open-pit mine planning. His research interests involve deterministic optimisation of large-scale problems and robust optimisation of uncertain problems.