Loading...
Design, development and implementation of operational fleet management systems using adaptative artificial intelligence techniques
Zamalloa Llerena, Lee Jornet
Zamalloa Llerena, Lee Jornet
Citations
Altmetric:
Advisor
Editor
Date
Date Issued
2024
Date Submitted
Collections
Research Projects
Organizational Units
Journal Issue
Embargo Expires
2025-11-26
Abstract
A fleet management system is a set of computational routines programmed and integrated with the main purpose of solving the truck-shovel allocation problem (TSA) and in consequence controlling the dispatch system of a fleet. In the mining industry, the fleet management system is responsible for calculating, deciding, and registering the best possible arrangement of trucks, shovels and other service equipment to achieve the production targets while meeting the operational constraints.
Due to the nature of the problem, the large quantity of live data involved, the uniqueness of each mining site, the computational limitations, and the uncertainties related to the mining activity, the existing research work and commercial software are unable to adjust the solution to the probabilistic variations in the calculation parameters, relying on a person called a dispatch operator for most of the parameters’ adjusting and decision-making process. This human agent is ultimately responsible for ensuring that the fleet resources are used in the best possible way. However, since the dispatch operator is a human being, the decisions may result in suboptimal solutions due to human-related factors such as experience, expertise, and communication skills, among others.
To provide a solution for this problem, a novel methodology is proposed in this thesis, which discusses a different solution process that involves the combination of Operations Research and machine learning techniques to evaluate best decision parameters and predict optimal decision-making for a variety of uncertain scenarios, providing the dispatch operator an optimal course of action for a set of operational conditions and a potential outcome for a given decision made.
To date, there is no algorithm, either from previous research work or commercially available in any software package, that can offer a reliable solution for these variable scenarios. Therefore, the set of steps, logic and reasoning described in this methodology represent a milestone in the analysis of the truck-shovel allocation problem through the integration of artificial intelligence techniques in the mining systems. Ultimately, the purpose of this research is to provide a smart tool to dispatch operators to compare and evaluate the accuracy of certain decisions in the continuous process of truck allocation and dispatching.
Associated Publications
Rights
Copyright of the original work is retained by the author.