Tuesday, September 25, 2018
Optimizing Truck Dispatching Decisions in Open-pit Mining using Integer Programming
A key operational problem in open-pit mining is the real-time dispatching of trucks between mining locations (where ore is loaded at trucks) and processing and waste ore sites (where ore is dumped). The goals in dispatching are consistently supply the processing sites with enough ore, to maintain ore quality at the processing sites (by mixing ore from different mining locations), and to meet extraction targets at the mining sites. The problem is challenging due to the potentially conflict between these goals and the need to make the dispatching decisions in real-time, in an environment with significant randomness in travel, loading, and unloading times, and in ore quality. We propose an optimization-driven approach to making dispatching decisions with a mixed-integer programming (MIP) model. We conduct a simulation study to compare the approach to two simpler policies that mimic industry practice, and find that the MIP-based dispatching approach significantly outperforms the competing policies on a wide variety of test cases, with the differences being most stark when the number of trucks is limited.
This is joint work with Jeff Linderoth and Amanda Smith from UW-Madison.
Learn more about Prof. James Luedtke.
Location: GB 202