Theses and Dissertations (Industrial and Systems Engineering)

Permanent URI for this collectionhttp://hdl.handle.net/2263/32358

Browse

Now showing 1 - 20 of 119

Supply chain resilience in a post-pandemic era : a scenario-based simulation study
Boshoff, Estella I. (University of Pretoria, 2026-02-17)
This study investigates supply chain resilience (SCR) through a systematic literature review and a simulation-based case study. Through a systematic literature review, the study aims to understand the definiton of SCR and its evolution, the different metrics proposed to measure SCR and interventions identified or proposed to improve SCR. After identifying the potential metrics and interventions, a case study is used to simulate and compare the effectiveness of a range of interventions on the resilience of a supply chain under disruption. As part of the literature review, a total of 1480 articles were reviewed with 233 considered relevant to this study. The findings highlight a shift in SCR research from risk management and robustness to adaptive, technology-driven approaches, particularly post-COVID-19. Various resilience metrics exist, including recovery time, financial impact, and network-based indices, though no single standard metric prevails. Interventions were categorised into redundancy, flexibility, network design, supplier strategies, Industry 4.0 and 5.0 strategies, recovery, and other measures. The case study considered a global supply chain with component manufacturing and product assembly based in China, exporting five products to the rest of the world. Distribution centres (DC’s) in Australia, South Africa, Russia, the USA, Brasil and Belgium serve customers around the world. Ten interventions from the flexibility, recovery, redundancy, supplier and network intervention categories were tested across 30 disruption scenarios. The comparison of many different interventions make this study unique as other simulation studies reviewed in the literature often test only one intervention. After analysing the results, it was determined that Dual Sourcing, a supplier intervention, is the most effective in improving service levels and reducing recovery time after disruption. Despite these insights, the study’s high-level approach presents limitations, including favourable assumptions about supplier recovery and disruption probabilities, limited intervention testing, and the exclusion of detailed manufacturing processes. Future research should explore lower-level interventions, such as additive manufacturing, and incorporate more granular simulation models to enhance practical applicability.
Integrated asset integrity and process safety analysis, management and holistic evaluation for sustainability of the onshore petrochemical installations
Ongwae, Benard Monte (University of Pretoria, 2026-02-18)
Significant accident hazards persist in onshore refining and petrochemical operations despite the widespread adoption of Asset Integrity Management (AIM) and Process Safety Management (PSM) systems. A central cause is that AIM and PSM are often implemented in parallel yet organizationally fragmented, whereas the human and cultural drivers of barrier performance remain underspecified and poorly governed. This thesis develops, validates, and operationalises an integrated AIM–PSM (iAIPSM) framework, grounded in competence, culture, and data, tailored for refineries and petrochemical facilities. The framework strengthens governance, embeds behaviour-first work aids at the point of risk, and connects frontline signals to enterprise data for timely, accountable decision-making, thereby advancing sector-relevant SDGs 3, 8, 9, and 13.The thesis addresses the fragmentation between AIM and PSM and the limited quantification of human and organisational factors in incident likelihood. It asks (i) the gaps that persist in literature and practice, (ii) how competency and process safety culture can be quantified and embedded into risk management models, (iii) how the competency–culture nexus links to frontline deviations such as cognitive overload, procedural drift under pressure, and (iv) how to consolidate diverse operational data into actionable leading indicators and governance dashboards. The objectives of this thesis are to (1) define an integrated AIM and PSM governance model; (2) quantitatively embed human and organisational‑culture modifiers within a fuzzy‑augmented bow‑tie framework; and (3) operationalise behaviour‑based controls such as cardinal rules, visual checklists, and frontline key performance indicators (KPIs) with demonstrated evidence of effectiveness.A mixed strategy was employed: (1) a PRISMA guided systematic review to establish the evidence base and specify the iAIPSM architecture; (2) a fuzzy augmented bow tie formulation that embeds competency profiles and organizational process safety culture as probabilistic modifiers of event frequencies and consequence pathways; and (3) a behavioural intervention and case study evaluation using concise cardinal rules, visual checklists, and KPIs mapped to human factor contributors and integrated into governance dashboards. Data integration and quality-assurance considerations are defined to connect point-of-work signals to enterprise systems for near-real-time oversight. The thesis demonstrates three central findings that collectively advanced the integration of asset integrity and process safety in onshore petrochemical operations. First, the proposed iAIPSM framework establishes a coherent model of integrated governance by clarifying the overlap between AIM and PSM, aligning organisational roles, and harmonising assurance activities, such as verification and management review. This unified structure is supported by a consolidated dashboard that links barrier health to behavioural compliance and highlights contributions to relevant Sustainable Development Goals, thereby strengthening both operational oversight and strategic alignment with sustainability. Secondly, the research shows that human and cultural factors can be quantified in a defensible and decision‑relevant manner. By applying fuzzy sets to represent equipment condition, workforce competence, procedural adherence, and organisational process safety culture, the study demonstrates how these modifiers materially influence risk estimates. In an illustrative loss‑of‑primary‑containment scenario, the baseline initiating frequency of 1.68 × 10⁻² yr⁻¹ is reduced to approximately 3.09 × 10⁻⁵ yr⁻¹ for the fire‑and‑explosion outcome once right‑hand‑side mitigations are applied. This result underscores the measurable sensitivity of event likelihood to human and cultural performance, offering a more holistic and auditable alternative to purely technical risk models.Finally, the thesis validates the effectiveness of behaviour‑centred controls through a turnaround case study. Behaviour‑first interventions, particularly the introduction of eight cardinal rules and visual checklists at the point of work, were associated with substantial reductions in critical procedural deviations, typically in the range of 80–93%. These improvements were consistently observed across high‑risk activities, including permit‑to‑work processes, lock‑out‑tag‑out isolations, and shift‑handover practices such as double‑block‑and‑bleed verification. Together, these findings demonstrate that integrating governance, quantifying human factors, and embedding behavioural controls can significantly strengthen process safety performance in complex industrial environments.The thesis makes three substantive contributions with direct implications for both industry practice and scholarly advancement. Firstly, it offers a practical, human‑centred integration of Asset Integrity Management and Process Safety Management, complete with explicit governance structures and assurance mechanisms that strengthen organisational accountability. Secondly, it establishes a quantitative bridge between workforce competencies, organisational process safety culture, and incident likelihood through a fuzzy‑augmented bow‑tie methodology, demonstrating how human and cultural dimensions can be incorporated into risk estimates in a transparent and defensible manner. Thirdly, it introduces a behavioural framework that operationalises governance intent in practice through cardinal rules, visual checklists, and targeted KPIs, ensuring that strategic expectations translate into consistent frontline execution. Collectively, these contributions from the iAIPSM construct, a transferable model that supports data‑driven oversight, capital prioritisation, and constructive regulatory engagement. The framework provides a coherent pathway toward zero‑incident operations in high‑hazard environments by unifying technical, human, and organisational levers within a single integrated system. The study is scoped to onshore midstream and downstream assets that focus on process hazard risk management. Data access constraints necessitated anonymisation and aggregation. Transferability to offshore or upstream settings requires caution and adaptation. The work is structured into seven chapters that guide the reader from a clear definition of the problem and its theoretical foundations through the development of the integrated framework and the application of quantitative modelling methods. It then validates the behavioural interventions and concludes with a synthesis of the study’s key insights and implications.
An economic production model for multi-product manufacturing with consideration towards complementary price-dependent demand, and the possibility of product recovery
Worrall, Kirsten (University of Pretoria, 2026-02-19)
In production industries, like stainless steel manufacturing, a shared resource can subsequently produce multiple complementary products that are consumed together. The production process results in defective outputs that can be recovered through reannealing and cutting coils into sheets on the constrained resource, while the customers’ demand for each product is impacted by the price of the complementary product. The study develops an Economic Production Quantity (EPQ) model for two mutually complementary products manufactured on a shared resource with imperfect production. Product Type A’s demand can only be satisfied through production from new feedstock, while Product Type B’s demand can be satisfied through remanufacturing non-conforming products and production from new feedstock. The model incorporates complementary price-dependent demands and remanufacturing to optimise lot sizes by minimising total cost, including setup, holding, rework and disposal. The two scenarios analysed are non-bottleneck and bottleneck scenarios, respectively. Numerical examples confirm global optimality and model feasibility. The proposed model provides an EPQ framework that can be used for industries to ensure cost efficiency and sustainability through integrated production planning.
Machine learning predictive model development for Coal Mine Methane (CMM) concentration detection in underground mining operations
Mooroogen, Rubeshen (University of Pretoria, 2025-12-03)
Despite underground mining being a major economic stay for several nations across the globe as a result of the need to explore diverse geo-resources ranging from coal, through oil and gas, to precious metals amongst others, for economic sustainability, emissions from the underground mining activities have posed a major challenge to the health of mine workers over the years. In a bid to narrow down, by focusing on coal mining, it could be said of the latter that it is a dangerous activity which has been notably responsible for large amounts of accidents resulting in the death of several mine workers globally. A key element that is responsible for the fatal aspect of underground coal mining is the presence and accumulation of toxic gases such as carbon monoxide, Sulphur, carbon dioxide and methane. The decision as well as the choice of coal mine methane (CMM) for this research is as a result of its susceptibility to explosions, fires and asphyxia. This is rendered possible as methane is a highly flammable gas as well as possessing the ability of displacing oxygen. This paper focused its investigation specifically on Coal Mine Methane (CMM) which is released as a result of the extraction of coal and the disturbance inflicted to surrounding rocks’ formation during deep mining operations. This research was based on the use of machine learning models to successfully predict dangerous concentrations of methane over the authorized threshold. Five machine learning classification models were implemented and compared with the objective towards finding the best model to predict and detect dangerous level of CMM. The models that were investigated include: Naïve-Bayes, logistic regression, Decision Trees, XG-Boost and artificial neural networks (ANN). Those predictions were made from a dataset containing information on the temperature, airflow, humidity, pressure and methane concentration at an underground coal mine. The temperature, airflow, humidity and pressure measurements were recorded by a series of sensors namely anemometers and component sensors. Furthermore, a vital research deliverable reached was the ability to evaluate and infer the most effective machine learning model for the research premised on a comparative investigation of five ML models. The results obtained using generated metrics such as the classification reports and the confusion matrices, resulted in the Artificial Neural Network (ANN) selection as the most effective ML technique. The summarized f1 scores for each model are described as follows: 0.28 for Naïve Bayes, 0.66 for the logistic regression, 0.42 for the decision tree, 0.23 for XG-Boost and 0.82 for the ANN. Furthermore, recommendation for building more robust machine learning models towards successfully predicting and detecting dangerous levels of CMM from an artificial intelligence’s perspective were provided.
Modelling for a sustainable coal-based direct reduction iron production using an integrated approach of circular economy and systems thinking for compliance and comprehensiveness and for eco-friendly environmental decision-making
Ncube, Ratidzo Yvonne (University of Pretoria, 2025-09-30)
The global shift towards sustainable industrial practices has intensified the need for environmentally responsible approaches to metallurgical processes, particularly coal-based Direct Reduction Iron (DRI). This study presents a holistic framework integrating the Hybrid Structural Interaction Matrix (HSIM), Circular Economy (CE) strategies, and System Dynamics (SD) to enhance the sustainability of coal-based DRI operations. This study identifies key waste drivers, maps their interactions, and quantifies the effects of 5R (Reduce, Reuse, Recycle, Recover, and Reclaim) CE interventions across multiple material and energy pathways. DRI metal is a preferred raw material to scrap metal in steel production; however, its production process is associated with substantial environmental burdens due to pollutants and waste generated. This research seeks to address the waste challenges and resource inefficiencies associated with the coal-based DRI process. Initially, HSIM was employed to prioritise and quantify material and energy flows within the DRI system, facilitating the identification of key waste streams. The Intensity Rating Factor (IRF) of the waste elements, computed from the HSIM, reveals that emissions and waste heat exhibited the highest weight intensity rating of 5.667, underscoring their substantial influence within the system. Subsequently, the prioritised streams are then aligned with the 5R framework (Reduce, Reuse, Recycle, Recover, and Reclaim) to develop targeted strategies for material circulation, thereby minimising losses and maximising resource efficiency. The Systems Thinking (ST) approach is utilised to generate a causal loop diagram, illustrating how waste from one stage of the process can be reused as input at another stage, thereby promoting material circularity. The developed 5R framework is then integrated into the DRI process to enhance resource efficiency, promote material recirculation, and minimise waste. A Circularity Index (CI) is derived as a metric to quantify the CE framework model for the coal-based DRI process. Case study data were utilised to illustrate the circularity index based on the 5R framework, resulting in a CI of 53% or 0.53. The CI value from the case study indicates a moderate level of circularity and significant potential for material utilisation, recycling, and energy recovery during production. Finally, a System Dynamics model is developed using VENSIM to simulate the temporal evolution of material reuse and recycling rates under various intervention scenarios. The integrated 5R framework is operationalised through dynamic modelling, enabling a predictive assessment of the process’s sustainability over a 12-month horizon. Data from secondary sources and several case studies were used to validate the System Dynamics model. Through sensitivity and scenario analyses, the model evaluates short-medium term impacts on waste reduction and energy recovery from various CE intervention levels: 10%, 30%, 50%, 70%, and 90%. Simulation results show substantial material reductions, especially for CE intervention levels above 50%. Virgin material demand decreases significantly at 90% intervention, resulting in more than 3.15 million tons of reduced material. Energy recovery almost doubles on the lower levels of CE intervention, levels of 10% to 50%CE intervention, reducing reliance on non-coking coal and lowering overall carbon dioxide emissions. The strategic implications of CE interventions at varying levels have many benefits, including a reduction in virgin material use, enhanced resource productivity, and lower fuel consumption.The HSIM-CE-SD approach developed in this research provides both a methodological innovation and a practical tool for sustainability improvement. By capturing both the structural complexity and dynamic behaviour of material interactions within the coal-based DRI system, this work bridges the gap between sustainability theory and industry practice in the metallurgical sector.
System dynamics and machine learning hybrid model for churn prediction in retail banking
Monyemonwa, Kakgiso Vanessa (University of Pretoria, 2026-01-28)
Customer churn in retail banking represents a dynamic, feedback-driven system in which customer satisfaction, service quality, acquisition, retention, and marketing interactions evolve over time. Existing churn prediction models predominantly rely on static machine learning approaches that fail to capture these endogenous churn dynamics. Resulting in models that have limited explanatory insight. This study develops and evaluates an integrated System Dynamics–Machine Learning (SD–ML) modelling framework to explain, simulate, and predict customer churn in a retail banking context. A system dynamics model was first constructed to represent the structural feedback mechanisms governing customer acquisition, retention, and attrition. The SD model puts into practice and tests the effects of service quality, customer satisfaction, marketing effectiveness, and word-of-mouth. At its core the SD model was used to mirror the complex, dynamic nature of customer churn. Quantify the interconnectedness and interrelationships of key churn variables within the system of churn. The SD model was validated and used to generate behavioural stock-and-flow trajectories, which are then extracted as dynamic features. These SD-generated features are integrated with source data, which included the demographic, financial, and campaign data. The dataset pipeline used to train and validate supervised machine learning classifiers entailed the SD-generated features and the preprocessed source dataset. The data preprocessing process included advanced feature engineering and feature selection. This was done extensively to enhance the predictive ability of the hybrid SD-ML model. The ML algorithms used in this study included XGBoost, CatBoost, and Random Forest. A total of three SD-ML models were built and compared using the standard metrics and the best-performing model. XGBoost was found to be the best-performing model based on the standard metrics, and was used to extract a forecast of customer churn, important feature variables and customer segmentation. Results demonstrate that incorporating SD-derived behavioural features improves churn trajectory forecasting and supports continuous customer segmentation based on churn risk and engagement patterns. The hybrid SD–ML model employing the XGBoost method demonstrated exceptional performance, with an accuracy of 99.80%, precision of 96.71%, recall of 97.05%, and an F1-score of 99.81%. An examination of feature importance showed that financial behaviour traits were the most important factors in predicting customer churn, accounting for 87% of the total predictive power. Customer segmentation revealed that segment one, which includes the young, low-income customers with significant credit risk and minimal subscription uptake, is the segment with the highest churn risk. This SD-ML hybrid uniquely models churn as a dynamic, feedback-driven system that combines the explanatory system insights with the data-driven prediction. The model provides an engineering decision-support tool for proactive churn detection, scenario testing, and targeted retention strategy design within complex service systems.
A Software-Selection-For-Process-Modelling (SSPM) tool for a tertiary educational setting
Baldwin, Matthew Gary (University of Pretoria, 2025-11-25)
Business process modelling plays a critical role in improving clarity, communication, and assessment in tertiary education. Yet selecting a usable, free-to-use modelling tool remains a persistent challenge for educators and students, particularly given inconsistent feature offerings, recurring usability problems, and the lack of comparative guidance in the literature. This study addresses that gap by designing and evaluating a Software-Selection-for-Process- Modelling (SSPM) artefact that supports transparent, requirement-driven tool choice in tertiary educational environments.Guided by Design Science Research, the study first confirmed a class of recurring usability problems by conducting a systematic review of peer-reviewed and grey literature and by experimenting with the commonly used Bizagi Modeler process modelling tool. The synthesis identified minimum and preferred requirements reflecting effectiveness, efficiency, satisfaction, learnability, and context of use. These requirements, together with evidence from the literature, formed the basis for constructing a comparative dataset of free-to-use tools. The SSPM uses this dataset through an integrated decision pipeline combining a structured preference-capture form with a retrieval-augmented generation (RAG) model and a multi-criteria decision analysis (MCDA) component. Each tool is represented as a structured semantic “chunk,” enabling similarity-based retrieval against user-specified requirements. Weighted MCDA scoring, controlled by an adjustable α-parameter using a geometric-mean calculation, integrates this retrieval output with documented tool evidence to generate ranked recommendations with full traceability.The SSPM was evaluated with two lecturers representing different tertiary contexts and through hands-on experimentation using Camunda, a tool identified through SSPM recommendations. Lecturer evaluations highlighted clear interface strengths, relevance of criteria, and transparency of the scoring logic, while also identifying practical integration limitations in the current prototype. With SUS scores improving substantially (45 to 77.5) following interface refinements. Experimentation confirmed that the SSPM’s predictions accurately reflected Camunda’s observed performance, validating the artefact’s ability to translate usability concerns into reliable tool recommendations.The study contributes: (1) a requirements set addressing the documented gaps in tertiary-oriented modelling-tool research; (2) a transparent SSPM artefact that combines RAG and MCDA for context-sensitive software selection; and (3) an evaluated demonstration showing practical decision-support value for tertiary educators. Future work may expand the SSPM to additional software domains, integrate enhanced usability and AI-related requirements, automate the retrieval workflow, and incorporate student-centred evaluation to explore tool suitability across broader tertiary use cases.
Economic order quantity model for growing items with seasonal dependent demand and deterioration
Takaedzwa, Ernest (University of Pretoria, 2026)
The agricultural sector is characterised by seasonal fluctuations in the demand for its products. To address these demand variations, innovative Economic Order Quantity (EOQ) models are required to effectively manage inventory for growing items. Growing items refer to living organisms bred for human consumption, such as livestock, fish, and poultry. This paper presents a novel lot-sizing model for growing inventory items, specifically tailored to chickens. The proposed model considers a sinusoidal time-varying demand pattern with the exclusion of stochastic factors. In addition, it accounts for the constant deterioration of the slaughtered inventory that occurs during the consumption phase. The proposed EOQ model is developed over a finite time horizon and assumes no lead time when purchasing broiler chicks and no mortality of the chickens during their growth phase. The primary objective of this model is to minimise total costs whilst optimising the number of production cycles. To achieve this, we employ a derived mathematical model in conjunction with a computational algorithm. To provide practical insights into the model’s effectiveness, a numerical analysis was conducted to evaluate the performance of key system indicators such as the total cost, the number of production cycles and the number of chicks to be purchased under various scenarios. The key numerical findings showed cost reductions and an optimal number of production cycles ranging from eight to nine cycles. Furthermore, the convexity of the objective function was examined to ensure its stability and reliability. A comprehensive sensitivity analysis was also performed, offering valuable managerial guidance regarding the model’s applicability and robustness in real-world applications. Notably, the key findings from the sensitivity analysis indicate that deterioration and the duration of the production period have the most significant impact on total system costs. This research contributes to the existing inventory management literature by extending new considerations to EOQ models for growing items. Understanding inventory management in the agricultural sector, particularly for products that experience seasonal demand and susceptibility to deterioration, is essential for optimising agricultural productivity. The EOQ model developed in this research enables farmers to reduce spoilage by up to 9.4% and thus increase the potential profits obtained from their farming operations. Such considerations within the farming sector have become increasingly critical in addressing the challenges associated with achieving consistent harvests and yields. Effective inventory management practices in the agricultural sector can mitigate losses, streamline supply chain processes, and enhance overall efficiency. These practices ultimately support the sustainability of agricultural operations within dynamic market environments.
Economic order quantity models for growing products with imperfect quality, storage constraints, and product interactions
Ladu, Micah (University of Pretoria, 2026-02-09)
In the area of inventory management and supply chains, there is a wide range of models that were developed to better simulate the real world. A recent significant breakthrough in inventory management that has become very popular lately, focusing on modelling inventory systems that handle products that can grow throughout the replenishment cycle. Examples of these items include livestock and poultry, which are essential products in the food industry, with the bulk of these products being sold downstream of most supply chains.In this research study, we consider growth as the product's ability to gain weight, which distinguishes the growing products from the conventional products. A common inventory system comprises two different phases: a growth phase, during which live newborn products are fed until a predetermined target weight is attained, and a consumption phase, which begins after slaughter, when the processed products are stored and gradually sold to meet constant demand. Feeding costs arise in order to feed the products throughout the growing phase, while a holding cost is applied for storing and maintaining the slaughtered products throughout the consumption phase. Within this framework, two lot sizing models for growing products are formulated for two distinct scenarios that may arise in food supply chains. To address these two cases, an Economic Order Quantity (EOQ) model for growing products was developed for each case. In addition to incorporating growing products into the EOQ model’s framework, the first model addresses the possibility that a portion of the growing products are imperfect (poor quality), potentially due to errors in the processing of the product and limited capacity in the growth and storage facilities. The second model incorporates the assumption that the two growing products are substitutable and that one of the products is made up of two complementary components. These models are intended to address the following critical concerns for inventory managers: What is the optimal order quantity? (i.e. How much to order?), What is the optimal replenishment cycle? (i.e. when to place an order?), What is the ideal point (weight and time) for the growing items to be slaughtered? And what order quantity is needed to substitute product 1 for product 2 and vice versa in stockout situations?The first model considers an inventory system in which a portion of the growing products are of poor quality, and that newborn products are purchased by the company. The newborn products are then raised until a predetermined target weight is attained, after which the products are slaughtered and processed. After slaughtering, the processed products then undergo a screening procedure where the good-quality products and the poor-quality products are separated before being put up for sale, and the poor-quality products are sold in bulk at a reduced amount. Another assumption is that the model further assumes that the company’s own growth and storage warehouse facilities have a limited amount of space, meaning, when the order quantity exceeded the total available space, additional external growth and storage facilities would have to be leased from external sources. Since the holding costs are greater at the rented facilities than the holding costs of the company’s own facilities, meaning products stored in the rented facilities were sold first to reduce total cost. In the second model, two growing items were considered, and it was assumed that one product can be substituted for another product due to product shortages, where one of the products consists of two components that are consumed together as complements. For each of the two models developed in this study, detailed descriptions are provided to the proposed inventory systems in order to guide the formulation of their corresponding mathematical models. Solution methodologies for the suggested mathematical models were provided as well. For each of the models proposed by the dissertation, numerical examples were also included to show the application of the solution techniques that were proposed. Additionally, for each of the models proposed by the dissertation, a sensitivity analysis was conducted on key parameters to show how changes in the inputs can affect the optimal order quantities and total profits of a company. Here are a few of the observations made based on the sensitivity analysis results. One of the major observations was that the existence of imperfect quality products within an inventory system indicates that additional newborn products must be procured to satisfy the required demand for good quality products. Which resulted in increased feeding costs, which would reduce the total profits by approximately 4.4%, and increased setup costs, which would reduce profits by approximately 6%. Capacity limitations in the growth and storage facilities raise the total costs and reduce total profits, mostly due to the rented facilities having a greater holding cost than the company-owned facilities. If the holding costs for the rented facility were increased, the total profits would be reduced by approximately 1.4%, and if the holding costs for the company’s facility were increased, the total profits would be reduced by approximately 7.1%. Although increasing the capacity of both the growth and storage company-owned facilities reduces the total cost, it involves enormous amounts of capital and financial risks in case the situation worsens in the market, at which point it would reduce the total profits by approximately 3.3%. The joint ordering policy for complementary and substitutable growing products shows that increased demand and usage rates increase the total profits by approximately 29.9% and 23.9% for the demand for products 1 and 2, respectively, while increasing the usage rates for product 1 increased profits by approximately 17.8% for a_1 and 35.7% for a_2, respectively. Conversely, increasing the setup costs for the products reduces total profits by approximately 2.0% and 1.2% for products 1 and 2, respectively. Increasing the holding and feeding costs for product 1 reduces total profits by approximately 3.2% and 2.9%, respectively, due to longer storage and feeding expenses but increasing the holding and feeding cost for products 2 barely affected the total profits because substitution shortens its effective storage duration. Increasing the weight of the slaughtered products reduces total profits by approximately 2.6% and 0.2% for products 1 and 2, respectively, due to the extended growth phase. Furthermore, an investigation was performed that shows that incorporating substitution has notable cost savings against classical EOQ models.Procurement, operation and inventory managers in industries dealing with growing products can use the inventory models outlined within this dissertation, which may provide some useful advice on how to deal with those products, specifically ordering decisions. By implementing these models, companies can greatly lower costs associated with inventory because growing products are key to food supply chains, and the resulting financial savings can be utilised in order to protect customer prices from inflation or increase total profits from a financial standpoint.
Interpretable deep reinforcement learning for hybrid flowshop scheduling under sequence dependent family setup time, unrelated machines and transportation time constraints
Nsionu, Gabriel Chukwuebuka (University of Pretoria, 2026-02-05)
Modern manufacturing environments require more than static production schedules: they demand adaptive and dynamic solutions capable of responding to frequent, irregular job orders and unforeseen events on the shop floor. To meet these challenges, schedules must be regenerated both rapidly and efficiently as new demands arise. This underscores the need for automated optimisation methods that can learn from experience, adapt to changing conditions, and update solutions at a pace that matches or exceeds the rate of change within the manufacturing system. This thesis explores the potentials of MatNet, a recent deep reinforcement learning (DRL) model, applied as a black box scheduler for generating high-quality schedules in realtime. While MatNet has shown promise in various combinatorial optimisation problems, its internal decision-making processes remain largely unexplored. This research aims to bridge that gap by systematically evaluating MatNet’s performance within an industrial scheduling context. A highly constrained hybrid flowshop scheduling problem with sequence dependent family setup times and transportation costs serves as a practical test case. To improve interpretability, Sobol sensitivity analysis is also employed to evaluate and provide insight into MatNets behaviour as a black-box scheduler. The results show that the MatNet-based DRL approach can effectively learn robust scheduling strategies, scale to larger problem instances beyond those seen during training and generate high-quality solutions within seconds. These capabilities demonstrate its suitability for deployment in real-world manufacturing environments where rapid response to unforeseen changes is critical
A software selection for event guest management method, using rule based reasoning
Mans, Christoph (University of Pretoria, 2025-12-09)
With the continual advancement in software development, the decision-making process for selecting appropriate software solutions becomes increasingly difficult. Literature indicates that the selection of incorrect software solutions has unfavourable consequences on the industry of interest. Within the event guest management (EGM) industry, a systematic literature search highlighted the need for a standard method on selecting an appropriate software application to support EGM. For this study, the requirements elicited, to support EGM, were not used to develop a new information system, but rather to select the most appropriate combination of software applications to support business EGM operations. The main research question was thus: What software selection for event guest-management method (SSEGM) can be amended/designed to compare event guest management tools to address the usability and efficiency needs of an events manager of large events? Literature was analysed on multiple industry and component software selection tools to identify decision-making techniques for selecting software which can be used for EGM. The rule-based reasoning (RBR) component within the hybrid knowledge-based system (HKBS) methodology was identified as a suitable decision-making method. Focusing on eliciting requirements, our scope of interest (SoI) was represented in a model for event guest-management (MEG) which was constructed from existing enterprise engineering body of knowledge concepts. In addition, requirements were extracted from existing EGM software applications by means of a grey literature review. A comprehensive set of requirements was populated into the RBR component using the design science research (DSR) methodology. The set of software requirements, and a database containing EGM software platforms and their respective features are thus included in the SSEGM. Multiple event managers at a large institution (LI) validated the completeness, relevance and usability of the SSEGM. We demonstrated the SSEGM in a simulation of an event at the LI to select a software solution capable of performing the required functions within the context of the LI’s event, linked to the EGM process. The selected software solution success was measured by determining whether it satisfies the event’s functional and non-functional requirements. Additionally, the study validated the SSEGM against its stated functional and non-functional requirements. Limitations were identified, which included method and SSEGM scalability, constructional and maintainability limitations. Future work, building onto the study, included incorporating case-based reasoning (CBR) and artificial intelligence components into the SSEGM and rebuilding the SSEGM in a low-code development platform. Recommendations are provided to promote the relevance, practicality and importance of the SSEGM. It was concluded that the SSEGM demonstrated its capability in guiding a user towards selecting an appropriate EGM software solution.
Six sigma design, analysis and evaluation of unscheduled bogie faults in freight wagons for optimum performance and enhanced logistics management
Photo, Tebogo Tlou (University of Pretoria, 2024-12)
In the South African economy, the rail industry is crucial in transporting large freight or consigning goods and humans at an affordable and safe cost. The rail industry's poor operational outlook and lack of competitiveness, caused by the frequent breakdowns of commuter trains, have become a significant challenge in recent years. This perceived level of unreliability, associated with the rail industry, has significantly impacted the downward trend of the competitiveness of trade and industry-related economic benchmarks. Improving the reliability of rolling stock in the rail industry has become a subject of great concern in many logistics and freighting organisations globally. Improving reliability implies better quality rolling stock will be available for running the trains. Improving rolling stock maintenance in the rail industry can be achieved through continuous supervision utilising effective business management techniques. Freight logistics companies need to be more proactive in the maintenance of their rolling stock. The complex competitive business challenges emanate from fluctuating market demands, globalisation, and economic uncertainty. The literature suggests that the adoption of DMAIC (Define-Measure-Analyse-Improve-Control) tools is a highly effective initiative that can help the rail industry compete with other sectors in the national economy by reducing operational costs, enhancing productivity, and improving overall quality. The rail industry, being a backbone industry, is the backbone of the South African economy and is significantly lacking in these areas. The purpose of this study was to enhance the overall performance of wagons through process optimisation in the sole rail logistics and freighting industry of the Republic of South Africa by exploring, understanding, and analysing the unscheduled bogie faults and identifying the factors that contribute to high number of faults using the Six Sigma DMAIC problem solving methodology. A case study was presented to validate the DMAIC framework. Twelve months of data, covering the 2023/2024 financial year for bogies faults, was collected and analysed. The study revealed that 2551 faults were recorded for twelve months. Using the Pareto chart, the findings showed that unscheduled bogie faults contributed 48% with a total of 618, followed by brakes with 460 faults and followed by hallow wear with 239 faults. Therefore, the findings disclosed a significant link between Six Sigma structural effectiveness and Six Sigma enablers. Furthermore, Six Sigma enablers play a pivotal role in enhancing organisational prosperity. This study recommends exploring Six Sigma flexibility, expanding the research scope, and investigating its application across diverse South African industries.
A system dynamics model of the relationship between investment decisions and a just renewable energy transition in developing countries
Munyadzwe, Desmond Basimane (University of Pretoria, 2024-09)
The adoption of renewable energy (RE) technologies in developing countries faces numerous barriers despite the urgent need for sustainable energy transition to combat climate change. This thesis investigates the dynamic interactions between investment decisions and renewable energy adoption in developing countries using a system dynamics modeling approach. The study specifically examines how investment in renewable energy influences the affordability and adoption rates of RE technologies, considering the complex interplay of economic, technological, and policy factors. By simulating various investment scenarios, this research highlights the critical role of equitable investment strategies in promoting a just energy transition. The model developed for this purpose, the Renewable Energy Affordability-Adoption System (REAAS), incorporates variables such as income levels, energy prices, and foreign financial aid to assess their impact on renewable energy affordability and adoption rates. The analysis identifies key leverage points where targeted interventions can significantly enhance the uptake of renewable energy. Findings suggest that increasing local investments, improving technology efficiency, and reducing dependency on energy imports are pivotal in making renewable energy more affordable and widely adopted in developing countries. Moreover, the study underscores the importance of balancing foreign and local investments to foster sustainable development and energy independence. This thesis contributes to the field by providing a nuanced understanding of how different investment strategies can affect renewable energy adoption in a context marked by economic and infrastructural limitations. The insights gained underscore the need for policies that support increased local capacity building and equitable distribution of renewable energy investments to ensure that developing countries can meet their energy needs sustainably.
Modelling, analysis and evaluation of a metric system for the quantification of management competitiveness in a system-of-systems complex human corporation
Schoeman, Maryke F. (University of Pretoria, 2024-11)
This research has quantified, through algorithmic sensing and metrication, the effective minimum management score required to attain competitiveness, otherwise known as the management index for competitiveness (MIC). The MIC is required by a System-of-Systems (SoSs) overseeing entity to competitively manage the complex network of systems that forms the heterogeneous SoSs cluster. Quantification of SoSs management is a scarcely researched field due to its extremely qualitative nature. This research has, however, presented an approach for quantifying the management effort. In a bid to bridge the research gap, a holistic and integrated framework depicting an SoSs network of 35 constituent systems in the agricultural grain industry was developed. The SoSs network was architected to show the complexities of the virtual and physical interactions between constituent systems. Furthermore, a quantitative mechanism via the Hybrid Structural Interaction Matrix (HSIM) concept was deployed. This was done so that instead of improving the overall management competitiveness through trial-and-error approaches, the priority systems can be identified and measured that will increase the overall competitiveness the most. From this, it was realised that the MIC herein is 0.50. The study also aimed to create specific rules that govern each level of competitiveness (by reflecting the necessary actions to be carried out and adhered to in order to maintain or enhance the competitiveness level). These governing rules are presented in a rules-like rubric format.
An approach towards process ownership within the financial sector
Erasmus, Maryka (University of Pretoria, 2024-11)
One of the main challenges when improving business processes, is the sustainability of the improvements. The purpose of this dissertation is to extract knowledge from existing research around process ownership (PO) as a hypothetical solution and re-structure it, using the enterprise engineering contextualization model (EECM). The intent is to develop a new approach for the deployment of PO, called Enablers Supporting effective Process Ownership (ESPO) that will enable both the design and management of PO as a new function at the enterprise. This study proposes a solution to the problem experienced at Co. INS, a short-term insurance company in South Africa where the main researcher is employed. People and/or information systems are frequently prioritized at the expense of processes. The hypothesis is that process ownership will address this misalignment between domains. The Primary Research Question is: “What enablers, supporting effective Process Ownership, will prioritize process management to ensure improved process performance and subsequently sustain a competitive advantage within the financial sector?”. The methodology followed is design science research (DSR) to develop and evaluate a first draft of the ESPO approach. A literature review was done initially, followed by thematic analysis. The synthesis which followed was done using EECM as reference model to construct the ESPO approach. The demonstration and evaluation of the artefact was done using small group discussions, questionnaires, and interviews. The literature confirmed the phenomena that a lack of contracted process ownership results in poor process performance with a negative effect on competitiveness. PO is dependent on governance, culture, and clear roles and responsibilities as key enablers. Existing literature, however, lacks a coherent approach to embed PO within a holistic business process management approach. EECM guides the development of the ESPO approach and elaborates on the why, what, and how the enterprise should evolve towards effective PO. During the demonstration and evaluation stages of ESPO at Co. INS, it was found that allocation of PO was done in some areas, but the effective deployment was lacking. As expected, the cultural assessments indicated various levels of readiness, correlated to the effectivity of PO. The importance of the process architecture and strategic alignment of processes were also highlighted as key components of the governance around PO. The demonstration and evaluation results indicate that ESPO effectively guides practitioners towards the achievement of sustainable process performance. Since PO has been confirmed as a critical capability for sustained process improvement, the ESPO approach is supporting the value of process as a strategic asset to create a sustained competitive advantage.
Integration of Multi-State Systems in a series of EPQ models for deteriorating products
Kapya, Tshinangi Fabrice (University of Pretoria, 2024-04-11)
This thesis contains a collection of problems dealing with the modelling and optimisation of multi-state production systems, hence addressing challenges within the broader category of flexible manufacturing. These systems are often subjected to random degradations, failures, age of machines, human errors, power supply disruptions, or changes in demand. In literature, many inventory models have been developed under the assumption that the lifetime of systems is infinite, meaning the performance of a system or equipment remains unchanged and is fully usable for satisfying future demand. Some other models have extended this assumption by considering the functioning of systems (or equipment) under binary modelling conditions in which two states are considered: operational state and failure state. However, a growing body of literature is beginning to take into consideration the numerous scenarios that may occur during the lifetime of an equipment. These situations contribute to the multiplicity of the possible states of systems. Such systems are called multi-state systems (MSS). MSS are generally subject to several failure modes, in particular degradation and age of the systems, with various effects on their performance. The operational characteristics of MSS allow them to continue to function; however, they have a reduced level of performance, demonstrating the adaptability and scalability of the equipment. In the literature, techniques to increase the performance of binary systems are often based on strategies including redundancy or preventive maintenance. In the case of multi-state systems (MSS), continuity of service is ensured by reconfiguration. The objective of this research is to develop models for managing inventory models for deteriorating items in a multi-state manufacturing environment. In many research based on the binary modelling conditions, ensuring the continuity of the production is an important issue. These models assume complete shutdowns of production systems upon failure of manufacturing resources, which can be extremely costly and lead to substantial manufacturing losses. By addressing these limitations that are present in many of the current literature, the models proposed in this thesis are more practical and thus beneficial for operations management practitioners when making decisions involving multi-state systems in manufacturing processes. For such systems, the breakdown or failure of any component only minimally or at least partially disrupts their performance. In this way, the system can continue to provide service with an acceptable level of degradation. The contribution of this thesis is the development of three mathematical models to optimise a series of Economic Production Quantity (EPQ) systems for deteriorating products. The first model deals with A lot-sizing model for a deteriorating product with shifting production rates, freshness-, price-, and stock-dependent demand with price discounting. The system consists of one machine producing a single type of product. When the component of the machine breaks down, the system is minimally or at least partially disrupted. Thus, it may continue to operate at a rate lower than the initial rate until a specific inventory level is reached. Initially, demand is influenced by its selling price and the level of stock displayed. As freshness declines, demand then depends on the product's freshness condition. As production continues, there is also a shift in production rate over time. To account for declining freshness affecting consumer interest and purchasing behaviour, discounts are applied after a certain period. The optimisation problem was solved using numerical methods and supported by sensitivity analysis to demonstrate its practical implications. However, at this stage, the model does not explore how raw materials with imperfect quality could impact this system. The second scenario presents a two-echelon supply chain inventory model for perishable products, incorporating a shifting production rate, stock-dependent demand rate, and imperfect quality raw material. This novel model extends the classic EPQ as well as the first novel developed in this thesis to account for the use of raw materials with imperfect quality in the production process. Two scenarios are formulated within this framework: one involves selling imperfect raw materials at a discounted price after a screening period, while the other entails keeping imperfect items in stock until they are returned to the supplier at the end of an inventory cycle. Both scenarios consider product deterioration as well as shifts in production rate. Numerical solutions were derived for these scenarios. The findings indicate that maximising profit may involve selling the proportion of imperfect raw material rather than retaining it until a new lot arrives from the supplier. This approach is particularly crucial in manufacturing systems where imperfect products appear in both the raw materials and finished goods. The results were validated through a sensitivity analysis. The third model expands previous novels by considering the scenario of a production system that continually declines, leading to an increasing rate of defects over time. It takes into consideration various elements including deterioration of finished products, stock levels, product quality, and the influence of corporate social responsibility (CSR). CSR plays a critical role in enhancing the reputation of the company, building customer loyalty, and increasing sales by demonstrating a commitment to ethical practices and societal well-being. The objective of the model presented in this scenario is to identify the optimal inventory level and cycle time that minimise the total cost per cycle. To illustrate the effectiveness of this model, numerical examples are provided along with sensitivity analysis. The findings show that the profit generated can increase by as much as 14 % if manufacturers integrate a setup cost policy and selling price decisions. Extending product shelf life by 60 % can increase the net profit by as much as 7 %. In another model involving a two-echelon supply chain system, the profit can be increased by as much as 360 % and 386 %, respectively, if the selling price and the demand enhancement parameter for inventory level increase by 20 %. Furthermore, the unit selling price can decrease the total cost by as much as 34 %. Operations managers can use all these mechanisms to increase profits in their production systems. Under reasonable conditions, other industrial fields like automotive, mineral processing plants, assembly lines, as well as the production of mechanical components, may also also benefit from the results obtained.
Heuristic solutions to minimise makespan in a hybrid flow shop scheduling environment with energy consumption constraints in steel making
Baloi, Brighton Miyelani (University of Pretoria, 2024-02-14)
Due to its advantage over other alternative energy efficient methods in cost savings, there has been an increasing interest in applying energy efficient scheduling in energy intensive industries faced with a conundrum of optimising production while consum ing less energy. In this dissertation, the energy-efficient hybrid flow shop scheduling problem is addressed to minimise the makespan without violating the total energy consumption threshold and where the threshold is not violated. None of the authors that apply the speed scaling mechanism under uniform parallel machines in the EHFSP domain consider each machine’s rate of response to a change in processing speed, the attack time. Therefore, the main contribution of this dissertation lies in addressing this gap,the consideration of attack time during speed scaling. The proposed algorithms seek to find the best makespan that incurs the minimum energy consumption where such alternative may exist. Energy is set as a constraint in re sponse to the dilemma faced by energy-intensive industries to continue fostering and sustaining competitive production by using less energy under an energy constraint. To solve the problem, two algorithms were proposed, each of which is an integration of some other scheduling heuristics, meta- and hyper-heuristics. The first is called the Improved Hyper heuristic NEH (IHNEH) algorithm, while the second is called the Improved Hyper heuristic GA (IHGA) algorithm. Each of the two algorithms operate in three stages, and share the first step, which is where the hyper heuristic is used to select a low-level heuristic for implementation in both solutions. The second step is what distinguishes the two solution. For the IHNEH algorithm, the NEH algorithm is used as the job sequencing procedure, while the IHGA makes use of the GA procedure. It was found that even though both algorithms were improved using the same improvement method, the IHNEH still displayed a superior performance over the IHGA especially for medium to large size problems in terms of the makespan and the energy consumption. The poor performance of the IHGA might be due to the random generation of the initial job sequence instead of using a constructive heuristic. The goodness of heuristic was used to measure the effectiveness of the two methods, and the computational results indicate that the methods were able to produce makespan values that deviate from the actual makespan by at most 2% for small size jobs. In terms of energy consumption, the IHNEH was able to produce energy consumption values that deviate from the actual energy consumption by 0% for small size jobs. For medium and large size jobs, the IHNEH had a deviation from the best makespan and energy consumption of 0, outperforming both the IHGA and the Branch and Bound under the time bound imposed. This implies that the IHNEH is a good technique for this problem given the energy threshold applied. Future studies can change that and study their influence on the Cmax and the energy consumption. Future work can also focus on extending the energy threshold further and study how many jobs can be processed without violating the threshold. The attack time values of machines were not based on actual industry data, however, future work could focus on obtaining these values to obtain more practical results. Also, speed scaling was not applied using actual machine speeds, but rather using speed factors, therefore future work could consider using actual speeds of machines. There are other alternative meta-heuristics that were not considered such as the PSO, and SA which are also capable of producing good results, and therefore they can also be used. Another suggestion for future research is initialising the GA with non random chromosomes to improve its performance. The performance of the GA is not only dependent on the quality of the initial solution and the termination criteria whose sensitivity analysis was presented, there are other factors that influence the GA such as the size of the population, the mutation and the crossover rate, therefore, it would be interesting to perform a sensitivity analysis of these parameters for future work.
Development of a lean six sigma framework for identification and minimisation of inefficiencies in construction projects.
Mophethe, Modiehi Mathabo Mirriam (University of Pretoria, 2024-02-15)
Introduction: This dissertation has focused on the systematic identification, analysis, and reduction of inefficiencies within the construction project domain through the utilisation of Lean Six Sigma. With the construction sector facing challenges such as budget overruns, delays, and resource misallocation, the need for effective strategies to enhance project efficiency becomes imperative. The Lean Six Sigma methodology, known for its success in various industries, offers a structured and data-driven approach to continuous improvement. This research adopts a comprehensive approach by incorporating literature reviews, case studies, and empirical data collection to explore the integration of Lean Six Sigma principles within the context of construction project management. The study begins by establishing the theoretical foundation, elucidating the fundamental concepts of Lean Six Sigma and its historical effectiveness in streamlining operational processes. Subsequently, it delves into a thorough evaluation of the unique constraints and complexities associated with construction projects, emphasising the importance of a tailored approach to enhance efficiency. Purpose: The objective of the dissertation is to carry out an extensive examination of the elements that lead to delays in projects and assess their subsequent effects. By comprehending the consequences and origins of these inefficiencies, adjustments were made to the lean six sigma tool to enhance the processes involved in construction projects and strive to reduce these inefficiencies to the greatest extent possible. As a result, both time and cost overruns were minimised, leading to savings in operational expenses. Consequently, the research endeavours to uncover the underlying causes of process inefficiencies and implement lean six sigma tools as effective solutions to address these inefficiencies. Approach: The practical aspect of the dissertation involves the utilisation of Lean Six-Sigma tools and methodologies in a real-life construction project, focusing on the identification of bottlenecks, waste, and unpredictability. To identify inefficiencies, a value stream map was constructed, while control charts were employed to measure variation. To gain a deeper understanding of the factors contributing to process inefficiencies, a fishbone diagram and factor analysis were utilised. Additionally, an EOQ model was employed to predict material lead time, which aids in effective planning. Furthermore, a scheduling and project monitoring tool, namely the CiteOps software, was developed, along with a visual dashboard created using PowerBi, enabling remote tracking and monitoring of project efficiency. The successful implementation of Lean Six Sigma in construction was illustrated through practical examples from previous studies, highlighting the adaptability and effectiveness of this framework. Findings: During the research, it was discovered that project delays and inefficiencies were largely influenced by lead time, delayed order placement, and task corrections. The correlation coefficient of lead time and placement was found to be 0.66, indicating a strong relationship between these factors and their significant impact on the efficiency of the project process. To mitigate long lead times, an EOQ model was implemented to forecast lead time and plan accordingly. The utilisation of project management software, such as CiteOps software, enhanced accountability among team members and facilitated better planning, enabling the timely identification and resolution of delays. By addressing these issues early on, their impact on the project was minimised. It is recommended to continue utilising the software and EOQ model to minimise the influence of factors that contribute to process inefficiency. Research limitations: The data used for the time study is only for the period when the project delays were at their climax and not from when they first occurred. This means the data used in this research may not be an accurate representation of the issue. The data sampled for this project may be insufficient due to limited availability. Originality: In this dissertation, the DMAIC approach was customised by combining the Six Sigma techniques, statistics, and differential equations to better quantify and understand the impact that delays have on the effective time spent on project completion. Keywords: Six Sigma, Lean, Lean Six Sigma, Inefficiencies, Inefficiency Minimisation, DMAIC
Dynamic system evaluation of fluctuating processor raw material on value chain financial sustainability
Ottermann, Helga (University of Pretoria, 2023)
As the global population grows, food self-sufficiency in developing countries becomes increasingly important and difficult (FAO, 2023). Due to variables like weather, agricultural production remains volatile, providing processors with an inconsistent supply of raw materials, making it difficult to operate at a reliable and sustainable utilisation and supply food consistently and competitively in the global market. This project aimed to evaluate the impact of a change in available raw material volume at the processing node on the total value chain’s financial sustainability. The project’s goal included illustrating and measuring the effect of seed availability, as well as determining the ideal amount of seed to be processed for maximum value chain financial sustainability. This was done with a system dynamic model that represented the Tanzanian sunflower value chain, including producers, traders and processors and measured financial sustainability with the net income indicator. Steps to develop the system dynamic model included the problem and background articulation, dynamic hypothesis with the causal loop diagram development, the mathematical formulation in AnyLogic, model verification and validation, and scenario analysis. The developed model represented the Tanzanian sunflower value chain accurately and assisted in gaining insight into managing different raw material availability disruptions (increasing and decreasing seed availability by 10% to 50%), which quantified the impact on each node’s financial sustainability with the net income indicator. This illustrated the efficiency gains due to economies of scale and supply and demand price trade-offs. Furthermore, a significant scientific contribution was to optimise the entire value chain to maximise the total value chain net income by determining the ideal amount of seed to be processed. The results illustrated how the model could contribute to quantifying a variety of different scenarios to analyse the impact of different interventions on the entire value chain system. The model quantified the financial sustainability of the value chain, however, as with most research, the model can be further improved to refine the results and incorporate different performance indicators (like environmental and governance) which may widen the reach of the results.
A sustainable operations management model in a non-profit organisation
Harmse, Martha Fredricka Petronella (University of Pretoria, 2023)
The sustainability of a non-profit organisation (NPO) in the South African education and research sector must be improved. Previously they had significant impact, but became under severe stress especially during COVID-19. This is an instance of NPOs in general whose sustainability is at risk. Although NPOs can improve their sustainability through operations management, the implementation of sustainable operations management requires further investigation. They can apply various models to improve the implementation of sustainable operations management, but a gap remains to develop such models. The purpose of this study is to develop an appropriate sustainable operations management model (SOMM) in the specific NPO. By following an action design research approach, the purpose of the study simultaneously is to develop a theory of how appropriate SOMMs can be developed in other NPOs. Most applications of action design research however involve information systems and technology. This study applies a less technologically orientated approach based on design research in education. Starting from a problem formulation phase, the actual problem is identified, conceptualised, and formulated as a case study that represents a class of research problems. Concepts are analysed through a literature review, and long-term commitment is obtained from the NPO. A building, intervention and evaluation phase starts with the contextualisation of a SOMM in the participating NPO, a research procedure is developed to address the actual problem, an interpretive framework and design ecology are developed to address the class of research problems, and effectiveness criteria are established. A SOMM is then iteratively developed through building, intervention and evaluation cycles until it is sufficiently refined. A reflection phase is executed in parallel with the previous two phases to capture the learning that occurs. Lastly, a formalisation phase addresses the reflexivity of the researcher and a design theory is formulated of how appropriate SOMMs can be developed in other NPOs. A practical contribution is made towards a SOMM in the NPO based on a definition of a model as a meta-theoretical framework to develop understanding, facilitate communication, propose improvements and to surface underlying assumptions. Sustainable operations management is defined as the management of human, natural, physical, financial and social capital and processes involved to satisfy self-defined needs and build resilience over the long term. The design starts by evaluating the current sustainability of the NPO, applies an integrated organisational perspective of a SOMM, regards sustainable operations management as an organised complex problem, and implements discordant pluralism. This entails organisational models for sustainability and systems thinking approaches namely a biomatrix entity systems perspective, viable system modelling, system dynamics, soft systems methodology, the Cynefin framework, and dynamic equilibrium modelling. The NPO confirms that the SOMM is effective in providing guidance to address their self-defined needs. These needs evolve through the development of the SOMM due to mutual influences between the model and the NPO. This ill-defined problem is addressed by changing the perceptions of the NPO to satisfy their needs, identify other needs, and to build resilience over the long term. The SOMM fosters and reinforces commitment to multiple, competing strategies by addressing paradox so that the NPO becomes more fluid, enhances their reflexive self-regulation through supportive capabilities, and becomes more sustainable. Design principles for the class of SOMMs in NPOs are based on the strategic selection of the case study, the interpretive framework, and the design ecology. A theoretical contribution is made towards sustainable operations management in NPOs in terms of key focus areas identified through content analysis of literature, and towards sustainable operations management in general with reference to the increasing number of hybrid organisations. The study also contributes to the theory of operations management modelling through the development of a research procedure to develop such a model. Furthermore, a contribution is made to a transformative research agenda of sustainability science in a design research mode. The study emphasises that enhanced sustainability does not imply predictability or a homeostatic balance to be achieved and maintained, but continuous tensions that must be creatively addressed. A less technologically orientated approach to action design research is proposed, and future research opportunities are identified.

Browse

Recent Submissions