Publication Type
PhD Dissertation
Version
publishedVersion
Publication Date
4-2024
Abstract
The precedence-constrained production scheduling problem (PCPSP) in Long-Term Mine Planning (LTMP) is NP-hard and conventionally prioritizes the Net Present Value (NPV) of profits. Even so, heightened sustainability concerns necessitate heightened sustainable practices. Yet, research still lags. This dissertation addresses this paucity by integrating sustainability elements through Multi-Objective Optimization (MOO), introducing novel algorithms and proposing an uncertainty assessment within a dual Multi-Objective Evolutionary Algorithm (MOEA) setup.
Firstly, our systematic review of past LTMP research focused on the PCPSP and highlighted sustainability elements. Overall, it furnished real-world components incorporated into mathematical formulations, trends, quality of solutions (efficacy) and computation time (efficiency) of various methods. These form the bedrock
later on to trade off the NPV of profits against environmental sustainability in a MOO. Particularly, we focused on the carbon dioxide emission costs (or carbon costs) which is the cost of absorbing carbon dioxide emitted during operations. With the generic PCPSP formulation, our MOO framework zoned into two approaches of decomposition-based and domination-based with their carbon costs formulations.
For the decomposition-based approach, we utilized a bounded objective function method and proposed a hybrid Temporally Decomposed Greedy Lagrangian Relaxation (TDGLR) algorithm. When evaluated against a Mixed Integer Programming (MIP) for a real-world operating mine, the TDGLR is faster and achieved minute gaps. For larger instances, the MIP failed to even provide feasible solutions. For the
domination-based approach, we leveraged two popular MOEAs of Non-dominated Sorting Genetic Algorithm II (NSGA-II) and Pareto Envelope-based Sorting Algorithm II (PESA-II). With NSGA-II, we illustrated the effectiveness of novel heuristics for the initial solution generation, crossover and mutation in forming an approximated Pareto front. Its solution sets were also diverse and close to that front. The front enables planners to adhere to stipulated annual carbon emission targets. Subsequently, the NSGA-II was compared to the PESA-II after experiments on the latter’s selection pressure parameter. PESA-II ran faster and its solution sets were more distributed. Meanwhile, NSGA-II converges better and steadily produced
non-dominated solutions. Moreover, we exemplified the threshold of ore tonnage deviations that maintains small alterations from the original results.
Finally, we surfaced several junctures for future studies. This comprise modifying the proposed MOEA framework to favor more complex datasets, including other sustainability elements (e.g. social) separately or concurrently, using stochastic means to measure uncertainty, and expanding to other uncertainties. Their considerations were also presented to further enable sustainable mining.
Keywords
Operations research, production scheduling, sustainability, multi-objective evolutionary algorithm, learning algorithms
Degree Awarded
Doctor of Engineering
Discipline
Artificial Intelligence and Robotics | Theory and Algorithms
Supervisor(s)
GUNAWAN, Aldy
First Page
1
Last Page
148
Publisher
Singapore Management University
City or Country
Singapore
Citation
AZHAR, Nurul Asyikeen Binte.
Enabling sustainable mining via AI-based techniques. (2024). 1-148.
Available at: https://ink.library.smu.edu.sg/etd_coll/587
Copyright Owner and License
Author
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.