Department of Industrial Engineering
Permanent URI for this community
Browse
Browsing Department of Industrial Engineering by browse.metadata.type "Masters"
Now showing 1 - 6 of 6
Results Per Page
Sort Options
- ItemCrafting asset allocation for a re-insurer via portfolio optimisation(Stellenbosch : Stellenbosch University, 2022-04) Abdulla, Mubeen; Von Leipzig, KonradENGLISH SUMMARY: One of the most challenging tasks faced by nancial advisors and consultants, relates to the phenomena of portfolio selection. This process typically entails selecting asset classes based on their risk and reward attributes. Striking an optimal balance between risk and reward is no easy task, given its con icting nature. This phenomena is referred to as portfolio optimisation and is commonly formulated and solved via the well-known mean-variance optimisation procedure, based on the pioneering works by Harry Markowitz. The objective function is formulated as a quadratic programming problem, that seeks to maximise expected return whilst minimising risk. While this approach presents an auspicious foundation to solve a portfolio optimisation problem, it does not incorporate the unique liabilities (such as future payments or claims) inherent to most institutional investors. The aim of the study is therefore to provide a roadmap outlining how assets and liabilities are dovetailed to enhance the decision making process around portfolio optimisation. To achieve this, the notion and premise of asset-liability management (ALM) and liability-driven investing (LDI) are introduced to better manage both assets and liabilities, coherently. This would ultimately ensure an institutional investor's long term nancial sustainability. To add a practical ingredient to this thesis, a real-world case study for a re-insurer is examined. Essentially, the roadmap is applied to a case study to solve a complete portfolio optimisation problem, from an LDI perspective. The results of the unconstrained asset allocation reveal the optimiser's preference to allocate chie y to a small range of asset classes. While this outcome may be theoretically appropriate, this presents a practical challenge given potential concentration risks, and lack of portfolio diversi cation opportunities. For this reason, constraints are imposed within the optimisation procedure, resulting in a more diversi ed and larger array of asset classes to include within a portfolio. To aid with the model validation component and to serve as credence, subject matter experts are consulted. The outcome of this validation was that the process embarked upon as well as the results produced are reasonable and resonates with industry standards. To supplement the model validation and to serve as a reasonability check, a comprehensive sensitivity analysis was undertaken on key input parameters such as expected return to assess the impact this has on the optimal portfolio of assets.
- ItemDevelopment of a digital rapid training course for improving the additive manufacturing adoption rate - fused filament fabrication(Stellenbosch : Stellenbosch University, 2022-04) van Wageningen, Roelof Pienaar; Hagedorn-Hansen, Devon; Von Leipzig, KonradENGLISH SUMMARY: Additive Manufacturing (AM) technologies, such as Fused Filament Fabrication (FFF), have a slow adoption rate. Training on these AM technologies is typically not included in primary to tertiary education curriculums and studies have shown that the lack of education on it, negatively affects the adoption rate. This issue was addressed in this study by developing a digital rapid training course on FFF. A literature study was first performed to gain a better understanding of the different AM technologies and the adoption thereof. The focus was then shifted to a set of learning methods and platforms that are used in the educational sphere. After completing the literature study, it was concluded that training users in FFF can help improve the adoption rate of the technology. The knowledge gained through the literature study was then used to develop a cross-platform digital training course (Web, iOS, and Android), aimed at introducing users to and educating them in FFF. The course consists of teaching sessions, tests, and questionnaires. The course was made available to the general public (free of charge) for a year with no specific target group, allowing users with and without FFF experience to participate. The training course automatically gathered quantitative and qualitative data by recording users’ answers during tests and questionnaires respectively. The course was completed by 198 participants. This data was then analysed to determine whether the training course increased the users’ knowledge of, confidence to engage with, and likelihood to adopt the FFF technology. From the group of participants, 87% claimed that their level of knowledge and understanding of FFF increased by participating in the course. The majority (94%) of the participants stated they are more likely to interact with the technology after participating. The users with no prior knowledge/experience with the technology were found to have benefited the most from the course. Such individuals can be targeted during the development and deployment of AM courses to have the biggest impact on the adoption rate. It was concluded that the training course increased the majority of users’ knowledge of, confidence to engage with, and likelihood to adopt the FFF technology.
- ItemDevelopment of a maintenance possession scheduler for a railway(Stellenbosch : Stellenbosch University, 2022-04) Cillie, Dewald; Bekker, James F.ENGLISH SUMMARY: Maintenance of rail infrastructure is an important element in rail operations in order to keep tra c moving. However, maintenance causes infrastructure to be taken out of service, which impacts tra c ow. In this study, the requirements of a maintenance possession scheduler for a South African application was investigated, and a proposed solution was subsequently developed. The main objective of the scheduler was to minimise the deviation of the train service on a subset of rail infrastructure while ensuring that the required maintenance is done. To achieve this, a literature study was done on a number of themes, which include an overview of the local railway operator with a look at the role of industrial engineering as a function in the railway operator business, railway infrastructure and operations, planning of railway operations, and maintenance in the context of rail operations. The topic of possession scheduling was then studied; the previous themes helped the researcher to learn the bigger picture while understanding possession scheduling is critical for this study. Past and recent works were studied and research areas and trends were synthesised, including time span of possession scheduling in optimisation models, and whether it was done on microscopic, mesoscopic or macroscopic level. The various optimisation objectives formulated by researchers were also noted, among other subthemes. An application case was identi ed as the railway infrastructure between Bellville and Wellington in the Western Cape province of South Africa. A novel mixed-integer linear programming model was formulated for this case and implemented in Cplex, after which it was validated. The model can do possession scheduling for 24 hours on a microscopic level. Finally, several experiments were conducted to investigate the performance and results of the model. It was found that the model delivered optimal results in less than eight minutes, which makes it a feasible maintenance possession scheduler for day-to-day work in the immediate planning horizon.
- ItemDevelopment of a smart trap for the surveillance of invasive fruit flies using internet of things and artificial intelligence(Stellenbosch : Stellenbosch University, 2022-04) Deacon, Quintus; Louw, Louis; Palm, DanielENGLISH SUMMARY: Invasive fruit flies are of major concern to the agricultural industry, causing millions of rands lost due to harvest damage, trade bans, and surveillance cost. Current surveillance methods of invasive fruit flies consist of entomologists manually inspecting fruit fly traps to determine the species of fruit flies captured. This process is time intensive, expensive, and inaccurate. This study proposes a smart trap approach based on vision system technology to automate the fruit fly species classification aspect of the surveillance process. The goal of the smart trap is to serve as an early warning system of invasive fruit fly outbreaks in pest free areas. A design science methodology was followed to design a smart trap that uses a camera imbedded in traditional fruit fly bucket traps to take images of new fruit fly captures and send them to a central server. Otsu's thresholding image segmentation was compared to the EfficienDet DO object detector for segmenting fruit fly instances from the image provided by the smart trap camera. EfficeintDet DO had the highest precision, recall, and Intersection over Union of 92%, 96.88% and 90.5% respectively. Thereafter pretrained models of EfficientNet BO, MobileNet V2, and MobileNet V3 Large were trained to differentiate between the Ceratitis capitata and -quilici fruit fly species segments provided by EfficientDet DO. MobileNet V3 Large had the highest accuracy and Fl-Score of 96.55% and 96.57% respectively. The object detection and image classification algorithms were trained on Google Colab using transfer learning and image augmentation. These were then executed on a Raspberry Pi 4 Model B microcomputer. The smart trap system was accurate in distinguishing between two fruit fly species, and capable of execution on a resource constrained device. The smart trap system shows promise for low cost, easy deployment smart traps but has some issues regarding connectivity in remote areas.
- ItemSimulating the South African forestry supply chain(Stellenbosch : Stellenbosch University, 2022-04) Laubscher, Jennifer Mignonne; Bekker, James F.; Ackerman, SimonENGLISH SUMMARY: The purpose of this study was to provide base simulation models that demonstrated some capabilities of discrete-event stochastic simulation when applied to the South African pulp- and saw timber supply chains from nursery- to mill-gate. The simulation models were designed to provide support for strategic decision-making by allowing scenario analysis through experimentation and bi-objective optimisation. A literature study was performed on various topics relating to the South African forestry industry, supply chain management and simulation modelling. This literature study included a discussion on various simulation software packages, from which one, namely Siemens Tecnomatix Plant Simulation was selected to be utilised in this study. To simplify the simulation modelling process two case studies were developed through stakeholder and subject-matter expert consultation. The case studies were converted into process ow diagrams, which could further be expanded into concept models. Thereafter, the simulation input data requirements were identi ed and the concept models could be translated into computerised versions in the chosen simulation software. The simulation models were built following a systematic and iterative approach. Final veri cation was done through structured model walk-throughs, error elimination and by performing entity tracing tests. Validation was done in consultation with project stakeholders and subject-matter experts, as well as through validation experimentation. After the simulation models were veri ed, validated and proven to be credible representations of the real-world pulpwood- and saw timber supply chains from nursery- to mill-gate, their experimental capabilities were demonstrated. The models were found to have a large combinatorial nature with regards to the number of experiments that can be performed, and exhibited the capability to perform \what-if" analysis and bi-objective optimisation.
- ItemTowards a framework to guide the development of ICT4D: A South African Perspective(Stellenbosch : Stellenbosch University, 2019-12) Coetzee, Lauren Lize; Grobbelaar, Sara; Bam, Wouter; Schutte, C. S. L.; Stellenbosch University. Faculty of Engineering. Dept. of Industrial Engineering.ENGLISH ABSTRACT: Information and Communication Technologies for Development (ICT4sD) has the capability of facilitating the flow of knowledge which offers developing countries opportunities to enhance systems that may assist in poverty alleviation and other developmental initiatives. Even though ICTs4D hold the potential of harnessing ICTs for social development, there still exists a significant percentage of ICT4D that fail to deliver the results they were developed for. Many ICT4D projects have made no difference and some have even caused harm in the communities they have been implemented. To determine how ICTs can be harnessed to serve as a catalyst and not as a hindrance for social transformation, one needs to consider the literature surrounding the topic of ICT4D. There is, however, an overall lack in the consistency of theory surrounding the process of development of ICT4D, and a lack of empirical evidence of appropriate methods regarding the development of ICT4D. To address this lack of evidence this study explores the best practices for the development of ICT4D in developing countries, specifically in South Africa (SA). The findings are used to develop a framework to guide the formation of ICT4D specifically within the Analysis and Design phases of development. An iterative seven-step process, by Jabareen (2011), the Conceptual Framework Analysis (CFA) process, was adopted to develop the ICT4D framework. These steps and thus the formation of the framework was accomplished within two overarching research phases, 1) a theoretical study and 2) an empirical study. The study is situated within three fields of literature, namely 1) Information System Development, 2) Human Development and 3) Information Communication Technologies for Development. The theoretical study investigated these three fields to understand the functioning of ICT4D from the perspective of all three of these fields and to develop a theoretical base of knowledge within each field. The exploration of these three fields formed part of the overview literature study and the Systematic Literature Review. The findings resulted in various analytical, design and functional concepts that were integrated to develop the preliminary Analytical and Design Framework. The empirical study followed by adopting a mixed methodology approach, comprising of two stages: 1) qualitative semi-structured interviews and (2) a quantitative framework-ranking exercise. The findings within this phase were applied to the framework to provide improvements and validation to develop the final framework. A positive response resulted from the empirical study and the framework was validated as needed, reliable, relevant and useful within the ICT4D domain. Even though the validity of the framework was established, further study is required to map the issues that may arise through implementation, and to confirm the usefulness thereof in real life situations. Since not one coherent approach exists for developing ICTs4D, but a combination of frameworks and tools are needed, the final framework is a contribution to the available tools and approaches that can be used to provide a guide to develop ICT4D within the Analysis and Design phases.