Doctoral Degrees (Industrial Engineering)

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 5 of 123
  • Item
    A decision support framework for the selection of appropriate time series forecasting methods in the retail sector
    (Stellenbosch : Stellenbosch University, 2023-10) Ganzevoort, Reinard Christiaan; Van Vuuren, Jan Harm ; Lindner, Berni G; Du Toit, Jacques; Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering.
    ENGLISH ABSTRACT:There is a significant trade-off in any high-turnover retail environment between limiting in-store inventory levels and mitigating the risk of stock-outs. This trade-off is caused by the typical aim of retail organisations to minimise the capital tied up in inventory without incurring a significant deterioration of their service levels (i.e. to ensure product availability for customers). In order, therefore, to better manage their inventories, retailers often consider the prediction of customer behaviour as a main priority. In practice, however, sales forecasting processes are usually automated to some extent and practitioners often have limited knowledge pertaining to the selection of appropriate forecasting methods. A generic framework is proposed in this dissertation for assisting retail forecasting practitioners in the selection of appropriate forecasting methods based on available time series data sets pertaining to retail sales. This forecasting framework takes as input a multivariate time series sales data set and facilitates the configuration, transformation and extraction of valuable information from these data in order to partition the data set into clusters of time series exhibiting similar attributes. The working of the framework is based on a generic, two-phased approach. One phase of the framework, called its benchmarking phase, involves establishing a benchmark data set (or updating it if it already exists) which can be leveraged to inform feature-based forecast model identification and ranking for different clusters of time series. The computationally efficient identification of a tailored shortlist of forecast models is thus facilitated during the other framework phase, called its implementation phase, for each sales time series presented to it by a retail organisation, based on the features of the time series presented. The two phases of the framework may be applied repeatedly in alternating fashion, thus enlarging the benchmark data set and improving its representativeness each time after having applied the implementation phase to the sales time series data of a new retail organisation. The framework is verified with reference to well-established retail sales benchmark data. The verified framework is employed to evaluate the difference in forecast quality and computational time, based on the benchmark data, that results from applying the forecasting methods recommended by the framework to newly presented retail timeseries data as opposed to exhaustively applying forecasting methods classified as traditional statistical techniques, machine learning techniques and ensemble techniques. The working of the framework is finally validated by applying computerised instantiations thereof to real-world data sets of time series representing retail sales.
  • Item
    A new simulation-based methodology for pro-active planning in deep-level mine ventilation systems to identify and mitigate hazards
    (Stellenbosch : Stellenbosch University, 2024-03) Jacobs, Daniël Rudolf; Schutte, Cornelius Stephanus Lodewyk; Van Laar, Jean Herman; Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering.
    ENGLISH ABSTRACT: Deep-level mining is present in various countries around the world. Gold production continues to decrease, and this places strain on the gold mining industry in South Africa. The depleting gold reserves meant that the existing deep-level gold mines had to expand deeper into the earth’s crust. Consequently, effective ventilation of deep-level mines is challenging. Deep-level mines rely on complex and dynamic ventilation systems to supply adequate air to underground workers. Changes to these systems are implemented to enable the expansion and deepening of the mines. These changes could cause certain hazards underground. The three main hazards that occur are high temperatures, gas accumulation and dust pick-up. It is therefore crucial to ensure that these hazards are prevented through effective planning. Digital twinning is a cutting-edge technology that simplifies the simulation and planning of the entire deep-level mine ventilation system. Currently, a problem persists in the absence of a concise strategy for identifying and mitigating hazards in life-of-mine planning, specifically when utilising a calibrated digital twin. Therefore, a systematic literature review was conducted to confirm this unique research opportunity. Additionally, a need is identified to determine the frequency of planning. There are currently two planning methods, namely incremental and end-state planning. The case study research methodology was utilised to develop a new strategy that uses a calibrated digital twin to identify and mitigate hazards in the planning of the life-of-mine. The strategy will then be verified in two parts, firstly verifying the strategy itself and secondly by utilising it in the two mentioned planning methods. The first verification case study implemented the hazard identification and mitigation strategy on a deep-level gold mine. The study produced a calibrated model with an accuracy of 95%. The calibrated model was then expanded according to the strategy and was used to identify various problem areas where high temperatures and insufficient airflow were present. These hazards were then mitigated, and sufficient ventilation was supplied throughout the three-year life-of-mine plan. The second verification case study implemented the hazard identification and mitigation strategy in both the incremental and end-state planning method. This enabled the comparison of these planning methods to evaluate the impact of the lower frequency of planning. This study highlights the significance of effective planning to minimise delays to ensure continuously safe working environments during the entire life-of-mine, rather than just at specific stages in the life-of-mine plan. Therefore, the developed solution in this research study can be used as a new simulation-based methodology for pro-active planning in deep-level mine ventilation systems to identify and mitigate hazards. The original contributions of the study include: • The development of a new strategy used in deep-level mine ventilation system life-of-mine planning. • The utilisation of a calibrated digital twin to identify possible problem areas in life-of-mine planning. • The reproducibility of the implementation of the strategy on all deep-level mines. • The improvement in the management and planning of a deep-level mine ventilation system. • The identification of which planning method is applicable for various applications.
  • Item
    Evolutionary algorithms for routing problems with drones with interceptions
    (Stellenbosch : Stellenbosch University, 2024-02) Ernst, Rudolf Heinrich; Groblem, Jacomine; Kaminsky, Phil; Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering.
    ENGLISH ABSTRACT: The process of fulfilling the final step of the delivery of people or goods from a fixed distribution facility is known as last-mile freight. Studies have shown that the bulk of costs related to logistics are related to ‘last mile’ distribution. The biggest loss in efficiency is observed in the final step of the delivery process. The ever-growing e-commerce business-to-customer industry has drastically increased the movement of goods in pursuit of satisfying the increased demand of individuals and retailers. Recent advances in technology have led to drones being used to make last-mile deliveries. The research in this dissertation contributes to showing the advantages of using drones in combination with trucks for deliveries. In this dissertation, a particle swarm optimization-based algorithm, a differential evolution-based algorithm, and a covariance matrix adaptation evolution strategy (CMA-ES) algorithm are developed and implemented on the travelling salesman problem with drone with interceptions (TSPDi). The metaheuristics are implemented on 23 benchmark datasets, varying in size and distribution of nodes. Two repair mechanisms are implemented to deal with infeasible solutions, a common problem observed when metaheuristics making use of continuous decision variable values are used to solve NP-hard combinatorial optimisation problems. The study showed that the self-adaptive neighbourhood search differential evolution (SaNSDE) algorithm is the best performer on smaller datasets, while the CMA-ES algorithm is the best algorithm for larger datasets. Further analysis is done to understand the performance of various metaheuristics on the combinatorial optimisation problem as the metaheuristics attempt to improve on their best solution by navigating through the search space. The analysis considers duplicate solutions, infeasible nodes, and infeasible solutions, as well as the efficiency of the metaheuristics when implemented on the TSPDi. The study highlights the problems encountered when metaheuristics are used to solve the combinatorial TSPDi. Four new SaNSDE algorithms are also developed for the TSPDi, focusing on utilising problem specific operators and heuristics to improve performance. The four new algorithms drastically improved the solutions obtained for the TSPDi, outperforming the originally developed metaheuristics by up to 80% on larger datasets. The newly developed SaNSDE algorithms also outperform the best-known results obtained by Moremi [169], by up to 77%. The results showed that the nearest neighbour for initial solutions (NNHis) SaNSDE is the best algorithm for the TSPDi. The NNHis SaNSDE is adapted to be applied to a multiple truck and drone logistical problem. In the first instance, a cluster-first, route second approach is implemented to solve a multiple travelling salesman problem with drone with interceptions (mTSPDi). A vehicle routing problem with drones with interceptions (VRPDi) is also solved where nodes are not clustered. The NNHis SaNSDE is expanded on by developing problem-specific operators to improve performance, and the newly developed NNHis EXDSH SaNSDE algorithm is applied to the VRPDi. Overall, the NNHis aNSDE mTSPDi approach performs the best, while the NNHis EXDSH SaNSDE improves upon the solutions obtained by the NNHis SaNSDE on the VRPDi.
  • Item
    Natural language processing for characterising the COVID-19 infodemic on South African twitter.
    (Stellenbosch : Stellenbosch University, 2024-03) Strydom, Irene Francesca; Grobler, Jacomine; Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering.
    ENGLISH ABSTRACT: The novel coronavirus disease of 2019 (COVID-19) was first detected in the city of Wuhan, China, and quickly spread to countries all around the globe. On the advice of the World Health Organization (WHO), governments-imposed lockdowns, social distancing, mask mandates, and other preventive measures that completely disrupted the daily lives of billions of people. Along with the disruption of daily life came fear, confusion, and anxiety, as news about the virus began circulating. Despite the attempts of the WHO and national governments to provide accurate information about the virus and prevent panic, rumours about its origin, effects, and cures surfaced on websites and social media. COVID-19 rumours became so prominent during the height of the pandemic that their spread became known as an “infodemic” and social media has been identified as a major contributing factor. The COVID-19 pandemic has exposed the potential harm that can be caused by misinformation and disinformation that is spread on social media. Scholars have responded by analysing content on social media to identify different kinds of misleading information about COVID-19 and to quantify how far it has spread. These studies make use of automated machine learning (ML) and natural language processing (NLP) techniques to analyse the large amounts of data present on social media. South Africa has, unfortunately, escaped neither the pandemic nor the infodemic. The full extent of the infodemic on South African social media, in contrast with other countries, is still unknown. ML and NLP techniques provide an opportunity to address this gap in research and characterise mis-/disinformation on South African social media. In this dissertation, two approaches were followed to characterise misleading information on South African Twitter. The first is a supervised ML approach that made use of a combination of transformer-based embedding models and feedforward neural network classifiers. The models were trained, optimised, and evaluated on publicly available, labelled COVID-19 Twitter misinformation datasets. The best performing model, LAMBERT, was then applied to unlabelled South African Tweets about the COVID-19 pandemic. Although the model performed well on the labelled test data (obtaining an F1–score of 89.9%), the model failed to reliably distinguish between mis-/disinformation Tweets and general Tweets in the unlabelled South African dataset. The second approach made use of an unsupervised topic modelling algorithm, BERTopic, to divide the unlabelled South African Tweets into coherent topics. The BERTopic model was trained and optimised on the unlabelled South African Tweets and produced 34 topics. By inspecting the representative terms and Tweets assigned to each topic, instances of mis-/disinformation were identified. The unsupervised approach was then refined by defining three novel procedures, namely discrete dynamic topic modelling (DDTM), topic evolution network formation (TENF), and topic characterisation (TC), to model the development of topics over time and characterise the extracted topics in terms of their textual, spatial, temporal and community facets. Using these procedures, networks of topics (including mis-/disinformation topics) were identified in the collected Twitter data. Lastly, these procedures were abstracted and combined to form a novel, generalised topic characterisation framework. This dissertation presents the first large-scale analysis of South African Twitter specifically aimed at characterising and mapping information disorder in the context of COVID-19, helping to better define the information disorder landscape on social media in the Global South and South Africa, in particular. The results described in the dissertation are a valubale departure point for future research and the proposed framework provides a comprehensive, yet flexible guide to characterising large corpora of text for domain experts and researchers alike.
  • Item
    A distributed simulation-optimisation system in support of goal pursuit in large-scale urban growth scenarios
    (Stellenbosch : Stellenbosch University, 2024-03) Van Heerden, Quintin; Van Vuuren, JH; Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering.
    ENGLISH ABSTRACT: Cities are complex systems and become increasingly complex as they grow. Urbanisation has to be managed carefully so as to avoid exerting unnecessary pressure on infrastructure and not to exacerbate further any unsustainable practices, such as overcrowding or urban sprawl. Urbanisation plays an important role in achieving the sustainable development goals of the United Nations, but local planning practices have to be aligned with these goals in order to achieve them. Local planning is plagued by requirements in national directives and legislation that mandate several planning instruments, frameworks, and policies, but do not provide clear direction on how to achieve these often-grandiose goals. Moreover, municipalities are not required to test the feasibility or the potential effects of their plans before implementing them. This may be due, in part, to a lack of scientific tools capable of assisting planners in this regard. State-of-the art land-use models, which are available for this purpose, are often too complex, require large volumes of data and specialist expertise to execute them, are limited in their application, or are too involved in terms of the underlying process of setting up appropriate test scenarios. A novel, generic system for long-term land-use planning is proposed in this dissertation which combines the powerful modelling paradigm of integrated land-use transport models with optimisation algorithms in a simulation-optimisation setting. The system comprises four functional components which together facilitate the pursuit of goals in large-scale urban growth scenarios. These components are a data component, a simulation component, an optimisation component, and an interpretation component. The main objective of the data component is to guide a decision-maker systematically through the processes of data collection, curation, preparation, and storage. The simulation component facilitates the establishment and execution of an integrated land-use transport model, while urban development aspiration levels may be specified in the optimisation component, which is aimed at performing multi-objective optimisation in pursuit of these targets. The working of the optimisation component is based on the execution of a self-adaptive metaheuristic responsible for managing various perturbation operators and interacts with the simulation component. Finally, the interpretation component provides a structured approach towards the interpretation of the performance of the entire system, the performance of the metaheuristic, as well as the output results with a view to make informed decisions with respect to land-use planning. integrated land-use transport models are notorious for their long execution times and require vast amounts of data and resources. The system proposed in this dissertation, therefore, conforms to one of two possible distributed system designs in order to guide the decision-maker during the process of establishing and running such a model in a distributed manner — either making use of microservices or else implementing the system on a high-performance computing cluster. An analysis of the costs involved is also carried out so as to assist the decision-maker in selecting the most appropriate system design for his or her needs.