Spatial epidemiological investigation of sport and leisure injuries in Victoria, Australia
- Authors: Singh, Himalaya
- Date: 2018
- Type: Text , Thesis , PhD
- Full Text:
- Description: Sport and leisure injuries are recognised as a public health issue in Australia. Despite the many health benefits associated with sport and leisure participation, there is a risk of sustaining injury during participation. To keep Australia active, there is a critical need to prevent injury occurrence. Epidemiological investigations in sport and leisure injuries have been largely examined by grouping of sports, age groups, sex and level of play. In addition, intrinsic (person-level) factors have been considered, such as strength, flexibility or previous injury history. These factors may not be sufficient to identify injury burden or prevent an increase in injury incidences. In the broader injury literature (e.g., road traffic crashes or drowning), it is known that injuries often cluster within specific places (i.e., road intersections or bodies of water). These specific geographic locations may also relate to sport and leisure injuries (e.g., sports grounds or facilities). Similarly, population-level factors such as socio-economic status or cultural groups within an area could influence the types of sports and leisure activities people participate in and consequently, the injuries that occur. A review presented in this PhD thesis revealed that there is very limited sport and leisure injury epidemiological information from a geographical perspective. To address this gap, and determine whether there is a spatial pattern in sport/leisure injuries, the aim of this PhD was to examine the geospatial distribution of sport/leisure injury hospitalisations and their association with a broad range of social and economic characteristics. This thesis uses spatial epidemiological methods to answer questions such as ‘Where do sports and leisure injuries occur?’ and ‘In whom do sports/leisure injuries occur?’ The main chapters present the results of the application of spatial epidemiological methods to describe the problem, to test hypotheses and to explore associations with possible explanatory variables. The findings showed a significant variation across metropolitan, regional and rural areas in the pattern and clustering of injuries when examining different sports, age groups and other variables such as education level. A secondary aim of this thesis was to consider the dissemination of sport and injury epidemiological data. As emphasised in the literature, there is limited spatial epidemiological information available to decision-makers and key stakeholders. At best, descriptive maps might be included in a report or research paper. However, these are static and limited to the results that the author chooses to present. Therefore, an important output from this PhD is a web-GIS application that has been specifically built to enable the exploratory analysis of sport/leisure injuries in Victoria. Sport and leisure injury prevention strategies and policy development relies on information about where, when, to whom and how sport/leisure injuries occur. This thesis demonstrates that a spatial epidemiological approach is an important and novel way to address epidemiological questions from a geographical perspective.
- Description: Doctor of Philosophy
- Authors: Singh, Himalaya
- Date: 2018
- Type: Text , Thesis , PhD
- Full Text:
- Description: Sport and leisure injuries are recognised as a public health issue in Australia. Despite the many health benefits associated with sport and leisure participation, there is a risk of sustaining injury during participation. To keep Australia active, there is a critical need to prevent injury occurrence. Epidemiological investigations in sport and leisure injuries have been largely examined by grouping of sports, age groups, sex and level of play. In addition, intrinsic (person-level) factors have been considered, such as strength, flexibility or previous injury history. These factors may not be sufficient to identify injury burden or prevent an increase in injury incidences. In the broader injury literature (e.g., road traffic crashes or drowning), it is known that injuries often cluster within specific places (i.e., road intersections or bodies of water). These specific geographic locations may also relate to sport and leisure injuries (e.g., sports grounds or facilities). Similarly, population-level factors such as socio-economic status or cultural groups within an area could influence the types of sports and leisure activities people participate in and consequently, the injuries that occur. A review presented in this PhD thesis revealed that there is very limited sport and leisure injury epidemiological information from a geographical perspective. To address this gap, and determine whether there is a spatial pattern in sport/leisure injuries, the aim of this PhD was to examine the geospatial distribution of sport/leisure injury hospitalisations and their association with a broad range of social and economic characteristics. This thesis uses spatial epidemiological methods to answer questions such as ‘Where do sports and leisure injuries occur?’ and ‘In whom do sports/leisure injuries occur?’ The main chapters present the results of the application of spatial epidemiological methods to describe the problem, to test hypotheses and to explore associations with possible explanatory variables. The findings showed a significant variation across metropolitan, regional and rural areas in the pattern and clustering of injuries when examining different sports, age groups and other variables such as education level. A secondary aim of this thesis was to consider the dissemination of sport and injury epidemiological data. As emphasised in the literature, there is limited spatial epidemiological information available to decision-makers and key stakeholders. At best, descriptive maps might be included in a report or research paper. However, these are static and limited to the results that the author chooses to present. Therefore, an important output from this PhD is a web-GIS application that has been specifically built to enable the exploratory analysis of sport/leisure injuries in Victoria. Sport and leisure injury prevention strategies and policy development relies on information about where, when, to whom and how sport/leisure injuries occur. This thesis demonstrates that a spatial epidemiological approach is an important and novel way to address epidemiological questions from a geographical perspective.
- Description: Doctor of Philosophy
Social exclusion : contested meanings, policy and experience
- Authors: Brass, Kate
- Date: 2009
- Type: Thesis , PhD
- Full Text:
- Description: "This thesis takes a critical inquiry approach to the concept of social exclusion, which has become a prevalent feature of the contemporary social policy lexicon ... [the] study takes an innovative methodological approach in that it explores and makes visible the diverse and contradictory meanings that practitioners and community members hold for social exclusion, and gives those meanings equal stature with the views of academic researchers and policy makers, as they are represented in the scholarly research literature and associated policy documents."
- Description: Doctor of Philosophy
- Authors: Brass, Kate
- Date: 2009
- Type: Thesis , PhD
- Full Text:
- Description: "This thesis takes a critical inquiry approach to the concept of social exclusion, which has become a prevalent feature of the contemporary social policy lexicon ... [the] study takes an innovative methodological approach in that it explores and makes visible the diverse and contradictory meanings that practitioners and community members hold for social exclusion, and gives those meanings equal stature with the views of academic researchers and policy makers, as they are represented in the scholarly research literature and associated policy documents."
- Description: Doctor of Philosophy
Single- and multiobjective reinforcement learning in dynamic adversarial games
- Authors: Kurniawan, Budi
- Date: 2022
- Type: Text , Thesis , PhD
- Full Text:
- Description: This thesis uses reinforcement learning (RL) to address dynamic adversarial games in the context of air combat manoeuvring simulation. A sequential decision problem commonly encountered in the field of operations research, air combat manoeuvring simulation conventionally relied on agent programming methods that required significant domain knowledge to be manually encoded into the simulation environment. These methods are appropriate for determining the effectiveness of existing tactics in different simulated scenarios. However, in order to maximise the advantages provided by new technologies (such as autonomous aircraft), new tactics will need to be discovered. A proven technique for solving sequential decision problems, RL has the potential to discover these new tactics. This thesis explores four RL approaches—tabular, deep, discrete-to-deep and multiobjective— as mechanisms for discovering new behaviours in simulations of air combat manoeuvring. Itimplements and tests several methods for each approach and compares those methods in terms of the learning time, baseline and comparative performances, and implementation complexity. In addition to evaluating the utility of existing approaches to the specific task of air combat manoeuvring, this thesis proposes and investigates two novel methods, discrete-to-deep supervised policy learning (D2D-SPL) and discrete-to-deep supervised Q-value learning (D2D-SQL), which can be applied more generally. D2D-SPL and D2D-SQL offer the generalisability of deep RL at a cost closer to the tabular approach.
- Description: Doctor of Philosophy
- Authors: Kurniawan, Budi
- Date: 2022
- Type: Text , Thesis , PhD
- Full Text:
- Description: This thesis uses reinforcement learning (RL) to address dynamic adversarial games in the context of air combat manoeuvring simulation. A sequential decision problem commonly encountered in the field of operations research, air combat manoeuvring simulation conventionally relied on agent programming methods that required significant domain knowledge to be manually encoded into the simulation environment. These methods are appropriate for determining the effectiveness of existing tactics in different simulated scenarios. However, in order to maximise the advantages provided by new technologies (such as autonomous aircraft), new tactics will need to be discovered. A proven technique for solving sequential decision problems, RL has the potential to discover these new tactics. This thesis explores four RL approaches—tabular, deep, discrete-to-deep and multiobjective— as mechanisms for discovering new behaviours in simulations of air combat manoeuvring. Itimplements and tests several methods for each approach and compares those methods in terms of the learning time, baseline and comparative performances, and implementation complexity. In addition to evaluating the utility of existing approaches to the specific task of air combat manoeuvring, this thesis proposes and investigates two novel methods, discrete-to-deep supervised policy learning (D2D-SPL) and discrete-to-deep supervised Q-value learning (D2D-SQL), which can be applied more generally. D2D-SPL and D2D-SQL offer the generalisability of deep RL at a cost closer to the tabular approach.
- Description: Doctor of Philosophy
Shelter From the Storm. Webs of connectedness and entanglement in contemporary painting of the everyday
- Authors: Griffin, Tony
- Date: 2021
- Type: Text , Thesis , PhD
- Full Text:
- Description: Considering the often-overlooked objects in a single Australian suburban home through the practice of still life painting, this project examines the everyday as a means to understand our anxious times. By employing phenomenological approaches this project explores the theoretical understandings of the everyday and many recent iterations of “thing theory”. Through the practice of still life, its traditions, language and its contemporary currency, painting is employed as a means to understand this age of anxiety. My research closely observes, documents and presents the everyday objects in a single suburban Australian home in the early part of the twenty-first century. Ian Hodder’s theory of entanglement and the mutual dependency between humans and things that it proposes, is considered as a suitable tool for a contemporary visual art practitioner in creating new understandings of our domestic and broader world. Additionally, my research employs the innovative and challenging approaches to the familiar championed by the French writer Georges Perec. His work in revealing an understanding of our world through the exhaustively comprehensive and meticulous description of everyday things provides a structural basis for this project. Our homes are where we experience the everyday nature of our existence most keenly and my home is not unlike the homes of others. It is a place that shares those broader ideals considered to constitute a home, my shelter and my refuge. It is a worthy place in which to seek an understanding of our complex world. This project reveals in paint my observations of the minor things which, when combined, constitute the major things in my small home. The result of these observations is a series of representations of a familiar environment that enables an audience to recognise their own surroundings and re-evaluate the many hidden entanglements in their world in more complex and evocative ways.
- Description: Doctor of Philosophy
- Authors: Griffin, Tony
- Date: 2021
- Type: Text , Thesis , PhD
- Full Text:
- Description: Considering the often-overlooked objects in a single Australian suburban home through the practice of still life painting, this project examines the everyday as a means to understand our anxious times. By employing phenomenological approaches this project explores the theoretical understandings of the everyday and many recent iterations of “thing theory”. Through the practice of still life, its traditions, language and its contemporary currency, painting is employed as a means to understand this age of anxiety. My research closely observes, documents and presents the everyday objects in a single suburban Australian home in the early part of the twenty-first century. Ian Hodder’s theory of entanglement and the mutual dependency between humans and things that it proposes, is considered as a suitable tool for a contemporary visual art practitioner in creating new understandings of our domestic and broader world. Additionally, my research employs the innovative and challenging approaches to the familiar championed by the French writer Georges Perec. His work in revealing an understanding of our world through the exhaustively comprehensive and meticulous description of everyday things provides a structural basis for this project. Our homes are where we experience the everyday nature of our existence most keenly and my home is not unlike the homes of others. It is a place that shares those broader ideals considered to constitute a home, my shelter and my refuge. It is a worthy place in which to seek an understanding of our complex world. This project reveals in paint my observations of the minor things which, when combined, constitute the major things in my small home. The result of these observations is a series of representations of a familiar environment that enables an audience to recognise their own surroundings and re-evaluate the many hidden entanglements in their world in more complex and evocative ways.
- Description: Doctor of Philosophy
Shared value priorities in groups : The impact of the Bahá’i Faith on Values
- Authors: Afshar, Faezeh
- Date: 2012
- Type: Text , Thesis , PhD
- Full Text:
- Description: Values such as honesty, meaning in life, pleasure, justice and freedom are guiding principles in motivating the selection and evaluation of individual behaviour and goals. They are also important for understanding cultural norms, attitudes and practices. In previous studies individuals and groups have often been found to be willing to subordinate their values toward goals and behaviours that are aligned with their own group; their ‘in-group’. Studies of values within a group have not reported evidence for the subordination of their values toward universal goals and behaviours; their ‘out-groups’. This research is the first to examine widely the values held by adherents of the Bahá’i Faith; a world-wide religious community that envisions a united global society. The main aim of this thesis is to discover and explain the system of value priorities held by Bahá’is. Data was obtained from an online survey of over one thousand responses to the cross-culturally validated ‘Schwartz Value Survey’. Data relating to the degree of commitment to religion referred to as religiosity, was also collected in addition to demographic information.
- Description: Doctor of Philosophy
- Authors: Afshar, Faezeh
- Date: 2012
- Type: Text , Thesis , PhD
- Full Text:
- Description: Values such as honesty, meaning in life, pleasure, justice and freedom are guiding principles in motivating the selection and evaluation of individual behaviour and goals. They are also important for understanding cultural norms, attitudes and practices. In previous studies individuals and groups have often been found to be willing to subordinate their values toward goals and behaviours that are aligned with their own group; their ‘in-group’. Studies of values within a group have not reported evidence for the subordination of their values toward universal goals and behaviours; their ‘out-groups’. This research is the first to examine widely the values held by adherents of the Bahá’i Faith; a world-wide religious community that envisions a united global society. The main aim of this thesis is to discover and explain the system of value priorities held by Bahá’is. Data was obtained from an online survey of over one thousand responses to the cross-culturally validated ‘Schwartz Value Survey’. Data relating to the degree of commitment to religion referred to as religiosity, was also collected in addition to demographic information.
- Description: Doctor of Philosophy
Semantic manipulation and business context in big data analytics
- Authors: Dinh, Loan
- Date: 2018
- Type: Text , Thesis , PhD
- Full Text:
- Description: Business organisations receive a huge amount of data from many sources every day. These data are known as big data. Since they are mostly unstructured, big data creates a complex problem of how to capture, manage, analyse and then derive meaningful information from them. To deal with the challenges that big data has brought, this research proposes a new technique in big data analytics in the business area to integrate semantically meaningful information relevant to textual queries and business context. To achieve this aim, this study makes three major related contributions. Firstly, the relationship between business processes and strategies is established using the concept of a rule-based inference model via facts and annotations. This relationship is required to determine the importance of a big data query for a business organisation. Secondly, we introduce approaches to determine the significance level of a query, by incorporating the processstrategy relationship, process contributions and priority of business strategies. Thirdly, the proposed data analytic technique embeds business context into the bedrock of data collection and analysis process. The first two contributions were implemented using Python programming language including the Pyke package (Pyke is built in the Python environment and has an artificial intelligence tool for the development of expert systems) and their performances were analysed based on a business use case. The last contribution was implemented mainly in the Hadoop and Java programs. Results show that the first contribution successfully establishes the processstrategy relationship, the second calculates the significance level of a query in relation to a business organisation, while the third reveals the huge impact of query significance level and business context on big data collection and captures deep business insights.
- Description: Doctor of Philosophy
- Authors: Dinh, Loan
- Date: 2018
- Type: Text , Thesis , PhD
- Full Text:
- Description: Business organisations receive a huge amount of data from many sources every day. These data are known as big data. Since they are mostly unstructured, big data creates a complex problem of how to capture, manage, analyse and then derive meaningful information from them. To deal with the challenges that big data has brought, this research proposes a new technique in big data analytics in the business area to integrate semantically meaningful information relevant to textual queries and business context. To achieve this aim, this study makes three major related contributions. Firstly, the relationship between business processes and strategies is established using the concept of a rule-based inference model via facts and annotations. This relationship is required to determine the importance of a big data query for a business organisation. Secondly, we introduce approaches to determine the significance level of a query, by incorporating the processstrategy relationship, process contributions and priority of business strategies. Thirdly, the proposed data analytic technique embeds business context into the bedrock of data collection and analysis process. The first two contributions were implemented using Python programming language including the Pyke package (Pyke is built in the Python environment and has an artificial intelligence tool for the development of expert systems) and their performances were analysed based on a business use case. The last contribution was implemented mainly in the Hadoop and Java programs. Results show that the first contribution successfully establishes the processstrategy relationship, the second calculates the significance level of a query in relation to a business organisation, while the third reveals the huge impact of query significance level and business context on big data collection and captures deep business insights.
- Description: Doctor of Philosophy
- Authors: McCoy, Jennifer
- Date: 2023
- Type: Text , Thesis , PhD
- Full Text: false
- Description: This thesis explores the contribution of Scottish selectors to the social development of the Omeo region of remote eastern Victoria, Australia, during the period 1875 to 1900, drawing on the experiences of people with familial connection to the McCoy family. These people were mostly agricultural labourers, prevented from land ownership in Scotland, and handloom weavers whose skills had been lost to the industrial revolution. Their lives had been changed by economic circumstances in Scotland, and by taking advantage of the Victorian Selection Acts, they achieved economic security while contributing to the development of the area. This thesis positions land as central to the economic and social development of this area. It was a remote, geographically isolated area where communication and climate challenges had defeated many. Using a documentary historical methodology, it examines the impact of each wave of Europeans, explorers, squatters, miners, and selectors on the land, with the corresponding dispossession of Aboriginal people, revealing the agency of those Aboriginal people in their responses, as well as the resilience of the settlers. Land records and demographic data have been used extensively to chart the land holdings and businesses of selected Scottish family members, positioning their place in the developing society, as community centres and activities flourished in response to a secure and growing population. As part of the Scottish diaspora in Australia, it highlights some of the qualities they brought with them to their new environment; it demonstrates the resilience people could bring to challenging circumstances; and it reveals the wealth of creative opportunities that grew and were encouraged, when limited communication to the outside world demanded a call on their own resources. The study also uncovers questions for further research, amongst them: a more detailed investigation of the role of newspapers, which are the primary source of historical information available; and the role played by women in building and supporting families even through crises, and whose records are buried in personal family records or diminished in newspaper stories.
- Description: Doctor of Philosophy
Seeds, soils and moisture : ecophysiology to inform mine site restoration in arid zones
- Authors: Duncan, Corrine
- Date: 2019
- Type: Text , Thesis , PhD
- Full Text:
- Description: Mining in arid regions of Australia is followed by restoration and rehabilitation efforts. However, attempts to reintroduce many woody plant species have been unsuccessful. Water is the key limiting resource to plant growth and seed germination in arid zones. In this thesis, I investigated ecophysiological processes that may help improve recruitment across novel arid landscapes. I tested seed traits and dormancy cues of eight keystone plants and found that six of them had non-dormant, rapidly germinating seeds; a trait we propose is an adaptation to the region’s unpredictable rainfall. To identify the germination niche of species, I incubated seeds under different temperatures and water potentials, and found drought avoidance to be an important survival strategy for arid species. I collected soil samples to compare biophysicochemical properties of reconstructed soils to remnant ecosystems, and found that the distribution of clay content in the reconstructed soils did not mirror the remnant soils, compromising their ability to sustain perennial vegetation. I also monitored soil moisture and found that soil reconstruction reduces rainfall infiltration and retention, and subsequently increases evaporation. The synthesis of these results demonstrate some of the limitations to successful restoration in these systems, such as (1) unknown dormancy cues and poor seed longevity, (2) infrequent and episodic plant recruitment due to water limitation, and (3) reduced hydrological function of reconstructed soils. The failure to reinstate hydrological function is the major constraint to ecological restoration in this arid zone. Nonetheless, results from this study suggest that restoration is possible through more strategic use of seed, careful selection of drought tolerant species, and increasing soil moisture. Further failures to reinstate ecosystem function and community dynamics in arid zones with reconstructed soils can be prevented by understanding the edaphic constraints to plant establishment, and ameliorating conditions to mimic ecohydrological processes in remnant ecosystems.
- Description: Doctor of Philosophy
- Authors: Duncan, Corrine
- Date: 2019
- Type: Text , Thesis , PhD
- Full Text:
- Description: Mining in arid regions of Australia is followed by restoration and rehabilitation efforts. However, attempts to reintroduce many woody plant species have been unsuccessful. Water is the key limiting resource to plant growth and seed germination in arid zones. In this thesis, I investigated ecophysiological processes that may help improve recruitment across novel arid landscapes. I tested seed traits and dormancy cues of eight keystone plants and found that six of them had non-dormant, rapidly germinating seeds; a trait we propose is an adaptation to the region’s unpredictable rainfall. To identify the germination niche of species, I incubated seeds under different temperatures and water potentials, and found drought avoidance to be an important survival strategy for arid species. I collected soil samples to compare biophysicochemical properties of reconstructed soils to remnant ecosystems, and found that the distribution of clay content in the reconstructed soils did not mirror the remnant soils, compromising their ability to sustain perennial vegetation. I also monitored soil moisture and found that soil reconstruction reduces rainfall infiltration and retention, and subsequently increases evaporation. The synthesis of these results demonstrate some of the limitations to successful restoration in these systems, such as (1) unknown dormancy cues and poor seed longevity, (2) infrequent and episodic plant recruitment due to water limitation, and (3) reduced hydrological function of reconstructed soils. The failure to reinstate hydrological function is the major constraint to ecological restoration in this arid zone. Nonetheless, results from this study suggest that restoration is possible through more strategic use of seed, careful selection of drought tolerant species, and increasing soil moisture. Further failures to reinstate ecosystem function and community dynamics in arid zones with reconstructed soils can be prevented by understanding the edaphic constraints to plant establishment, and ameliorating conditions to mimic ecohydrological processes in remnant ecosystems.
- Description: Doctor of Philosophy
Rule-based interactive assisted reinforcement learning
- Authors: Bignold, Adam
- Date: 2019
- Type: Text , Thesis , PhD
- Full Text:
- Description: Reinforcement Learning (RL) has seen increasing interest over the past few years, partially owing to breakthroughs in the digestion and application of external information. The use of external information results in improved learning speeds and solutions to more complex domains. This thesis, a collection of five key contributions, demonstrates that comparable performance gains to existing Interactive Reinforcement Learning methods can be achieved using less data, sourced during operation, and without prior verifcation and validation of the information's integrity. First, this thesis introduces Assisted Reinforcement Learning (ARL), a collective term referring to RL methods that utilise external information to leverage the learning process, and provides a non-exhaustive review of current ARL methods. Second, two advice delivery methods common in ARL, evaluative and informative, are compared through human trials. The comparison highlights how human engagement, accuracy of advice, agent performance, and advice utility differ between the two methods. Third, this thesis introduces simulated users as a methodology for testing and comparing ARL methods. Simulated users enable testing and comparing of ARL systems without costly and time-consuming human trials. While not a replacement for well-designed human trials, simulated users offer a cheap and robust approach to ARL design and comparison. Fourth, the concept of persistence is introduced to Interactive Reinforcement Learning. The retention and reuse of advice maximises utility and can lead to improved performance and reduced human demand. Finally, this thesis presents rule-based interactive RL, an iterative method for providing advice to an agent. Existing interactive RL methods rely on constant human supervision and evaluation, requiring a substantial commitment from the advice-giver. Rule-based advice can be provided proactively and be generalised over the state-space while remaining flexible enough to handle potentially inaccurate or irrelevant information. Ultimately, the thesis contributions are validated empirically and clearly show that rule-based advice signicantly reduces human guidance requirements while improving agent performance.
- Description: Doctor of Pholosophy
- Authors: Bignold, Adam
- Date: 2019
- Type: Text , Thesis , PhD
- Full Text:
- Description: Reinforcement Learning (RL) has seen increasing interest over the past few years, partially owing to breakthroughs in the digestion and application of external information. The use of external information results in improved learning speeds and solutions to more complex domains. This thesis, a collection of five key contributions, demonstrates that comparable performance gains to existing Interactive Reinforcement Learning methods can be achieved using less data, sourced during operation, and without prior verifcation and validation of the information's integrity. First, this thesis introduces Assisted Reinforcement Learning (ARL), a collective term referring to RL methods that utilise external information to leverage the learning process, and provides a non-exhaustive review of current ARL methods. Second, two advice delivery methods common in ARL, evaluative and informative, are compared through human trials. The comparison highlights how human engagement, accuracy of advice, agent performance, and advice utility differ between the two methods. Third, this thesis introduces simulated users as a methodology for testing and comparing ARL methods. Simulated users enable testing and comparing of ARL systems without costly and time-consuming human trials. While not a replacement for well-designed human trials, simulated users offer a cheap and robust approach to ARL design and comparison. Fourth, the concept of persistence is introduced to Interactive Reinforcement Learning. The retention and reuse of advice maximises utility and can lead to improved performance and reduced human demand. Finally, this thesis presents rule-based interactive RL, an iterative method for providing advice to an agent. Existing interactive RL methods rely on constant human supervision and evaluation, requiring a substantial commitment from the advice-giver. Rule-based advice can be provided proactively and be generalised over the state-space while remaining flexible enough to handle potentially inaccurate or irrelevant information. Ultimately, the thesis contributions are validated empirically and clearly show that rule-based advice signicantly reduces human guidance requirements while improving agent performance.
- Description: Doctor of Pholosophy
Robust Mobile Malware Detection
- Authors: Khoda, Mahbub
- Date: 2020
- Type: Text , Thesis , PhD
- Full Text:
- Description: The increasing popularity and use of smartphones and hand-held devices have made them the most popular target for malware attackers. Researchers have proposed machine learning-based models to automatically detect malware attacks on these devices. Since these models learn application behaviors solely from the extracted features, choosing an appropriate and meaningful feature set is one of the most crucial steps for designing an effective mobile malware detection system. There are four categories of features for mobile applications. Previous works have taken arbitrary combinations of these categories to design models, resulting in sub-optimal performance. This thesis systematically investigates the individual impact of these feature categories on mobile malware detection systems. Feature categories that complement each other are investigated and categories that add redundancy to the feature space (thereby degrading the performance) are analyzed. In the process, the combination of feature categories that provides the best detection results is identified. Ensuring reliability and robustness of the above-mentioned malware detection systems is of utmost importance as newer techniques to break down such systems continue to surface. Adversarial attack is one such evasive attack that can bypass a detection system by carefully morphing a malicious sample even though the sample was originally correctly identified by the same system. Self-crafted adversarial samples can be used to retrain a model to defend against such attacks. However, randomly using too many such samples, as is currently done in the literature, can further degrade detection performance. This work proposed two intelligent approaches to retrain a classifier through the intelligent selection of adversarial samples. The first approach adopts a distance-based scheme where the samples are chosen based on their distance from malware and benign cluster centers while the second selects the samples based on a probability measure derived from a kernel-based learning method. The second method achieved a 6% improvement in terms of accuracy. To ensure practical deployment of malware detection systems, it is necessary to keep the real-world data characteristics in mind. For example, the benign applications deployed in the market greatly outnumber malware applications. However, most studies have assumed a balanced data distribution. Also, techniques to handle imbalanced data in other domains cannot be applied directly to mobile malware detection since they generate synthetic samples with broken functionality, making them invalid. In this regard, this thesis introduces a novel synthetic over-sampling technique that ensures valid sample generation. This technique is subsequently combined with a dynamic cost function in the learning scheme that automatically adjusts minority class weight during model training which counters the bias towards the majority class and stabilizes the model. This hybrid method provided a 9% improvement in terms of F1-score. Aiming to design a robust malware detection system, this thesis extensively studies machine learning-based mobile malware detection in terms of best feature category combination, resilience against evasive attacks, and practical deployment of detection models. Given the increasing technological advancements in mobile and hand-held devices, this study will be very useful for designing robust cybersecurity systems to ensure safe usage of these devices.
- Description: Doctor of Philosophy
- Authors: Khoda, Mahbub
- Date: 2020
- Type: Text , Thesis , PhD
- Full Text:
- Description: The increasing popularity and use of smartphones and hand-held devices have made them the most popular target for malware attackers. Researchers have proposed machine learning-based models to automatically detect malware attacks on these devices. Since these models learn application behaviors solely from the extracted features, choosing an appropriate and meaningful feature set is one of the most crucial steps for designing an effective mobile malware detection system. There are four categories of features for mobile applications. Previous works have taken arbitrary combinations of these categories to design models, resulting in sub-optimal performance. This thesis systematically investigates the individual impact of these feature categories on mobile malware detection systems. Feature categories that complement each other are investigated and categories that add redundancy to the feature space (thereby degrading the performance) are analyzed. In the process, the combination of feature categories that provides the best detection results is identified. Ensuring reliability and robustness of the above-mentioned malware detection systems is of utmost importance as newer techniques to break down such systems continue to surface. Adversarial attack is one such evasive attack that can bypass a detection system by carefully morphing a malicious sample even though the sample was originally correctly identified by the same system. Self-crafted adversarial samples can be used to retrain a model to defend against such attacks. However, randomly using too many such samples, as is currently done in the literature, can further degrade detection performance. This work proposed two intelligent approaches to retrain a classifier through the intelligent selection of adversarial samples. The first approach adopts a distance-based scheme where the samples are chosen based on their distance from malware and benign cluster centers while the second selects the samples based on a probability measure derived from a kernel-based learning method. The second method achieved a 6% improvement in terms of accuracy. To ensure practical deployment of malware detection systems, it is necessary to keep the real-world data characteristics in mind. For example, the benign applications deployed in the market greatly outnumber malware applications. However, most studies have assumed a balanced data distribution. Also, techniques to handle imbalanced data in other domains cannot be applied directly to mobile malware detection since they generate synthetic samples with broken functionality, making them invalid. In this regard, this thesis introduces a novel synthetic over-sampling technique that ensures valid sample generation. This technique is subsequently combined with a dynamic cost function in the learning scheme that automatically adjusts minority class weight during model training which counters the bias towards the majority class and stabilizes the model. This hybrid method provided a 9% improvement in terms of F1-score. Aiming to design a robust malware detection system, this thesis extensively studies machine learning-based mobile malware detection in terms of best feature category combination, resilience against evasive attacks, and practical deployment of detection models. Given the increasing technological advancements in mobile and hand-held devices, this study will be very useful for designing robust cybersecurity systems to ensure safe usage of these devices.
- Description: Doctor of Philosophy
Revision and validation of the Australian Competency Standards for Occupational Therapy Driver Assessors to ensure best practice
- Authors: Fields, Sally
- Date: 2022
- Type: Text , Thesis , PhD
- Full Text:
- Description: Occupational therapy driver assessment and rehabilitation is an advanced and complex area of practice. Users of this service require that occupational therapy driver assessors are, and remain, competent in their role. Competency standards support and inform occupational therapy driver assessment and rehabilitation practice, ensure standardisation of practice, and determine whether practice can be deemed competent. Competency standards for occupational therapy driver assessors in one Australian state, Victoria, were published in 1998 to provide a benchmark for service provision and guide training. Since then, occupational therapy driver assessor practice has undergone further developments and growth, including a significant increase in research evidence underpinning this advanced practice area. As a result, the competency standards required revision to reflect contemporary practice and enable their use and application across all Australian states and territories. The purpose of the research program reported in this thesis was to revise and validate the Australian Competency Standards for Occupational Therapy Driver Assessors and determine if education and practice using these standards led to increased adherence to their use in clinical practice. A sequential, exploratory, mixed methods design was used within a pragmatist research paradigm to revise and validate the competency standards for occupational therapy driver assessors from a national perspective. Four separate research studies were conducted. In the first study, an overview was undertaken to summarise the current evidence for including cognitive and perceptual assessments within fitness-to-drive evaluations, to inform the revision of the competency standards for occupational therapy driver assessors. The second study involved focus groups with Australian occupational therapy driver assessors and expert educators and identified the changes and revisions required to be made to the Victorian competency standards for occupational therapy driver assessors. Content themes and changes were extracted from the focus groups and combined with research evidence to draft the revised standards. The third study collected feedback from consumers and Australian driver licensing authorities to incorporate into the revised draft. This was followed by a two-round Delphi process with Australian occupational therapy driver assessors to achieve consensus and validate the Australian Competency Standards for Occupational Therapy Driver Assessors. A high level of consensus on the content of all competency standards and underlying practice behaviours was obtained in the final round. Finally, a mixed-methods multiple case-study methodology was adopted, and multiple forms of evidence were collected from file audits, semi-structured interviews and an individualised education session with occupational therapy driver assessors. This study demonstrated that, for this sample, occupational therapy driver assessors’ adherence to the Australian Competency Standards for Occupational Therapy Driver Assessors increased following an individually tailored education package about the standards. The research in this thesis has contributed to the evidence base for occupational therapy driver assessor practice. The Australian Competency Standards for Occupational Therapy Driver Assessors have been revised, validated and published, and the impact of an education package to increase adherence has been investigated. These competency standards: (i) assist occupational therapy driver assessors’ clinical practice, support occupational therapy registration and regulatory agencies to evaluate practice against a standard and (iii) provide universities with a framework to deliver postgraduate training in occupational therapy driver assessment and rehabilitation. Finally, this research has described a process that can be utilised in other areas of occupational therapy practice to develop and validate competency standards.
- Description: Doctor of Philosophy
- Authors: Fields, Sally
- Date: 2022
- Type: Text , Thesis , PhD
- Full Text:
- Description: Occupational therapy driver assessment and rehabilitation is an advanced and complex area of practice. Users of this service require that occupational therapy driver assessors are, and remain, competent in their role. Competency standards support and inform occupational therapy driver assessment and rehabilitation practice, ensure standardisation of practice, and determine whether practice can be deemed competent. Competency standards for occupational therapy driver assessors in one Australian state, Victoria, were published in 1998 to provide a benchmark for service provision and guide training. Since then, occupational therapy driver assessor practice has undergone further developments and growth, including a significant increase in research evidence underpinning this advanced practice area. As a result, the competency standards required revision to reflect contemporary practice and enable their use and application across all Australian states and territories. The purpose of the research program reported in this thesis was to revise and validate the Australian Competency Standards for Occupational Therapy Driver Assessors and determine if education and practice using these standards led to increased adherence to their use in clinical practice. A sequential, exploratory, mixed methods design was used within a pragmatist research paradigm to revise and validate the competency standards for occupational therapy driver assessors from a national perspective. Four separate research studies were conducted. In the first study, an overview was undertaken to summarise the current evidence for including cognitive and perceptual assessments within fitness-to-drive evaluations, to inform the revision of the competency standards for occupational therapy driver assessors. The second study involved focus groups with Australian occupational therapy driver assessors and expert educators and identified the changes and revisions required to be made to the Victorian competency standards for occupational therapy driver assessors. Content themes and changes were extracted from the focus groups and combined with research evidence to draft the revised standards. The third study collected feedback from consumers and Australian driver licensing authorities to incorporate into the revised draft. This was followed by a two-round Delphi process with Australian occupational therapy driver assessors to achieve consensus and validate the Australian Competency Standards for Occupational Therapy Driver Assessors. A high level of consensus on the content of all competency standards and underlying practice behaviours was obtained in the final round. Finally, a mixed-methods multiple case-study methodology was adopted, and multiple forms of evidence were collected from file audits, semi-structured interviews and an individualised education session with occupational therapy driver assessors. This study demonstrated that, for this sample, occupational therapy driver assessors’ adherence to the Australian Competency Standards for Occupational Therapy Driver Assessors increased following an individually tailored education package about the standards. The research in this thesis has contributed to the evidence base for occupational therapy driver assessor practice. The Australian Competency Standards for Occupational Therapy Driver Assessors have been revised, validated and published, and the impact of an education package to increase adherence has been investigated. These competency standards: (i) assist occupational therapy driver assessors’ clinical practice, support occupational therapy registration and regulatory agencies to evaluate practice against a standard and (iii) provide universities with a framework to deliver postgraduate training in occupational therapy driver assessment and rehabilitation. Finally, this research has described a process that can be utilised in other areas of occupational therapy practice to develop and validate competency standards.
- Description: Doctor of Philosophy
Rethinking english language support : reflections from chinese international students on communicating orally in english in the study abroad context
- Authors: Pruis, Karen
- Date: 2019
- Type: Text , Thesis , PhD
- Full Text:
- Description: Australian universities are currently emphasizing such graduate qualities as English language proficiency. However, this can be problematic for many international students who may still be struggling to communicate orally in English while participating in their chosen course of study. Indeed, the English language proficiency of international students has often received negative attention from the Australian media, employers, and academics. Several government reviews have recommended ongoing English language support for international students that extends beyond existing English language entry requirements, English Academic Preparation programs, and Post-entry Language Assessments. At the same time, there has been a critique of responses to these issues which focus on language proficiency—and simply place the onus on international students to increase their English language proficiency—without taking issues of intercultural communicative competence into account, along with calls for institutional frameworks to better support student experience. This thesis is one contribution to this debate, supported and informed by an exploration of lived experiences of Chinese students engaging with oral English language use and other supports. Not only are Chinese students the largest cohort of international students in Australia, but English education in China also does not appear to prepare them effectively to communicate orally in English. Very different sociocultural backgrounds can also make it difficult for this cohort to engage both academically and socially in Australian higher education. These issues comprise road blockers that can limit English language practice opportunities and negatively affect academic success. iii This thesis shines a light on this cohort’s oral English language experiences in the study abroad context via engagement of Max van Manen’s (1990, 2014) frameworks for phenomenologically oriented qualitative research. Drawing on the lived experiences of Chinese students and a critical review of the regulatory frameworks that inform sectoral and institutional English language proficiency models, this thesis argues for an English language proficiency support framework that integrates insights from intercultural communicative competence into oral English language proficiency as well as the development of curriculum, course delivery, and assessments.
- Description: Doctor of Philosophy
- Authors: Pruis, Karen
- Date: 2019
- Type: Text , Thesis , PhD
- Full Text:
- Description: Australian universities are currently emphasizing such graduate qualities as English language proficiency. However, this can be problematic for many international students who may still be struggling to communicate orally in English while participating in their chosen course of study. Indeed, the English language proficiency of international students has often received negative attention from the Australian media, employers, and academics. Several government reviews have recommended ongoing English language support for international students that extends beyond existing English language entry requirements, English Academic Preparation programs, and Post-entry Language Assessments. At the same time, there has been a critique of responses to these issues which focus on language proficiency—and simply place the onus on international students to increase their English language proficiency—without taking issues of intercultural communicative competence into account, along with calls for institutional frameworks to better support student experience. This thesis is one contribution to this debate, supported and informed by an exploration of lived experiences of Chinese students engaging with oral English language use and other supports. Not only are Chinese students the largest cohort of international students in Australia, but English education in China also does not appear to prepare them effectively to communicate orally in English. Very different sociocultural backgrounds can also make it difficult for this cohort to engage both academically and socially in Australian higher education. These issues comprise road blockers that can limit English language practice opportunities and negatively affect academic success. iii This thesis shines a light on this cohort’s oral English language experiences in the study abroad context via engagement of Max van Manen’s (1990, 2014) frameworks for phenomenologically oriented qualitative research. Drawing on the lived experiences of Chinese students and a critical review of the regulatory frameworks that inform sectoral and institutional English language proficiency models, this thesis argues for an English language proficiency support framework that integrates insights from intercultural communicative competence into oral English language proficiency as well as the development of curriculum, course delivery, and assessments.
- Description: Doctor of Philosophy
Restoration ecology in the semi-arid woodlands of north-west Victoria
- Authors: Murdoch, Fiona
- Date: 2005
- Type: Text , Thesis , PhD
- Full Text:
- Description: Arid areas are often overgrazed and dysfunctional with poor recruitment of desirable species, diminished control over resources and altered soil properties. Restoration ecology re-establishes these valued processes. State-and-transition models summarise knowledge of vegetation dynamics and tools for restoration, and encourage the incorporation of new information. The model developed here for semi-arid woodlands of north-west Victoria highlighted the unknown cause of observed, natural recruitment and the need for a technique, other than direct seeding and handplanting, for enhancing the recruitment of desirable species. I pursued these knowledge gaps for two dominant, woodland trees: Allocasuarina luehmannii and Casuarina pauper. Natural recruitment of juvenile C. pauper was found to be limited and primarily from root suckers. Extensive recruitment of A. luehmannii was shown to be mostly seedlings established following substantial reductions in grazing pressure since 1996. Seedlings were associated with areas devoid of ground flora near a female tree. The importance of competition between seedlings and ground flora, spatial variation in soil moisture and individual variation in the quantity of seed produced deserves further investigation to enhance future restoration success. Root suckers of both C. pauper and A. luehmannii can be artificially initiated, albeit in low numbers and this was found to be a feasible, new tool for restoration. Suckers are preceded by the growth of callus tissue on exposed or damaged, living, shallow roots. Both male and female trees can produce suckers and spring treatments may be more successful. Genetic fingerprinting of mature A. luehmannii and C. pauper trees in six populations did not identify any clonal individuals indicating that recruitment in the past has been from seedlings. Despite this, the high level of gene flow suggests that the impact of introducing small numbers of root suckers into existing populations is unlikely to impact negatively on the population genetics of these species.
- Description: Doctor of Philosophy
- Authors: Murdoch, Fiona
- Date: 2005
- Type: Text , Thesis , PhD
- Full Text:
- Description: Arid areas are often overgrazed and dysfunctional with poor recruitment of desirable species, diminished control over resources and altered soil properties. Restoration ecology re-establishes these valued processes. State-and-transition models summarise knowledge of vegetation dynamics and tools for restoration, and encourage the incorporation of new information. The model developed here for semi-arid woodlands of north-west Victoria highlighted the unknown cause of observed, natural recruitment and the need for a technique, other than direct seeding and handplanting, for enhancing the recruitment of desirable species. I pursued these knowledge gaps for two dominant, woodland trees: Allocasuarina luehmannii and Casuarina pauper. Natural recruitment of juvenile C. pauper was found to be limited and primarily from root suckers. Extensive recruitment of A. luehmannii was shown to be mostly seedlings established following substantial reductions in grazing pressure since 1996. Seedlings were associated with areas devoid of ground flora near a female tree. The importance of competition between seedlings and ground flora, spatial variation in soil moisture and individual variation in the quantity of seed produced deserves further investigation to enhance future restoration success. Root suckers of both C. pauper and A. luehmannii can be artificially initiated, albeit in low numbers and this was found to be a feasible, new tool for restoration. Suckers are preceded by the growth of callus tissue on exposed or damaged, living, shallow roots. Both male and female trees can produce suckers and spring treatments may be more successful. Genetic fingerprinting of mature A. luehmannii and C. pauper trees in six populations did not identify any clonal individuals indicating that recruitment in the past has been from seedlings. Despite this, the high level of gene flow suggests that the impact of introducing small numbers of root suckers into existing populations is unlikely to impact negatively on the population genetics of these species.
- Description: Doctor of Philosophy
Resolving the diversity paradox : an analysis of the diversity-conflict-performance paradigm in an Australian context
- Authors: Qin, John
- Date: 2009
- Type: Text , Thesis , PhD
- Full Text:
- Description: "Despite the intensive efforts to measure and predict the effects of group diversity on performance, research has produced extremely inconsistent and mixed results. This state of knowledge has presented a diversity paradox suggesting coexisting and conflicting effects of diversity. In order to explain the paradox and therefore improve our understanding of diversity, a three-way relationship (i.e. diversity-conflict-performance identified as a paradigm) has been suggested as a promising explanation. This thesis explores the effects of diversity via the paradigm, thereby offering a deeper insight into the diversity paradox." --Abstract.
- Description: Doctor of Philosophy
- Authors: Qin, John
- Date: 2009
- Type: Text , Thesis , PhD
- Full Text:
- Description: "Despite the intensive efforts to measure and predict the effects of group diversity on performance, research has produced extremely inconsistent and mixed results. This state of knowledge has presented a diversity paradox suggesting coexisting and conflicting effects of diversity. In order to explain the paradox and therefore improve our understanding of diversity, a three-way relationship (i.e. diversity-conflict-performance identified as a paradigm) has been suggested as a promising explanation. This thesis explores the effects of diversity via the paradigm, thereby offering a deeper insight into the diversity paradox." --Abstract.
- Description: Doctor of Philosophy
Resistance training efficiency : The efficacy of paired sets
- Authors: Robbins, Daniel
- Date: 2009
- Type: Text , Thesis , PhD
- Full Text:
- Description: A paired set (PS) refers to the coupling of exercises targeting agonist-antagonist muscle groups and performed coincidentally. Although PS have been prescribed by strength and conditioning coaches as a means of developing strength and power, scientific research investigating their efficacy is limited. The aim of this research was to investigate not only the efficacy of PS, but also their efficiency.
- Description: Doctor of Philosophy
- Authors: Robbins, Daniel
- Date: 2009
- Type: Text , Thesis , PhD
- Full Text:
- Description: A paired set (PS) refers to the coupling of exercises targeting agonist-antagonist muscle groups and performed coincidentally. Although PS have been prescribed by strength and conditioning coaches as a means of developing strength and power, scientific research investigating their efficacy is limited. The aim of this research was to investigate not only the efficacy of PS, but also their efficiency.
- Description: Doctor of Philosophy
Research projects reports and professional and ethical issues report
- Authors: Molloy, Mari
- Date: 2004
- Type: Text , Thesis , PhD
- Full Text:
- Description: "The thematic link within each of the individual reports arising from three of these placements is that of anxiety and depression"
- Description: Doctor of Psychology (Clinical)
- Authors: Molloy, Mari
- Date: 2004
- Type: Text , Thesis , PhD
- Full Text:
- Description: "The thematic link within each of the individual reports arising from three of these placements is that of anxiety and depression"
- Description: Doctor of Psychology (Clinical)
Research projects 1, 2, 3 and 4
- Authors: Keogh, Melissa
- Date: 2002
- Type: Text , Thesis , PhD
- Full Text:
- Description: These projects deal with 2 different themes. Firstly the training of parents whose children suffer from DSM-IV attention-deficit /hyperactivity disorder. Secondly therapy for the treatment of separation disorder.
- Description: Doctor of Psychology (Clinical)
- Authors: Keogh, Melissa
- Date: 2002
- Type: Text , Thesis , PhD
- Full Text:
- Description: These projects deal with 2 different themes. Firstly the training of parents whose children suffer from DSM-IV attention-deficit /hyperactivity disorder. Secondly therapy for the treatment of separation disorder.
- Description: Doctor of Psychology (Clinical)
Reliability-based design optimisation methods in large scale systems
- Authors: Ezzati, Ghasem
- Date: 2015
- Type: Text , Thesis , PhD
- Full Text:
- Description: Doctor of Philosophy
- Description: Structural optimisation is an important field of applied mathematics, which has proved useful in engineering projects. Reliability-based design optimisation (RBDO) can be considered a branch of structural optimisation. Different RBDO approaches have been applied in real world problems (e.g. vehicle side impact model, short column design, etc.). Double-loop, single-loop, and decoupled approaches are three categories in RBDO. This research focuses on double-loop approaches, which consider reliability analysis problems in their inner loops and design optimisation calculations in their outer loops. In recent decades, double-loop approaches have been studied and modified in order to improve their stability and efficiency, but many shortcomings still remain, particularly regarding reliability analysis methods. This thesis will concentrate on development of new reliability analysis methods that can be applied to solve RBDO problems. As a local optimisation algorithm, the conjugate gradient method will be adopted. Furthermore, a new method will be introduced to solve a reliability analysis problem in the polar space. The reliability analysis problem must be transformed into an unconstrained optimisation problem before solving in the polar space. Two methods will be introduced here and their stability and efficiency will be compared with the existing methods via numerical experiments. Next, we consider applications of RBDO models to electricity networks. Most of the current optimisation models of these networks are categorised as deterministic design optimisation models. A probabilistic constraint is introduced in this thesis for electricity networks. For this purpose, a performance function must be defined for a network in order to define safety and failure conditions. Then, new non-deterministic design optimisation models will be formulated for electricity networks by using the mentioned probabilistic constraint. These models are designed to keep failure probability of the network below a predetermined and accepted safety level.
- Authors: Ezzati, Ghasem
- Date: 2015
- Type: Text , Thesis , PhD
- Full Text:
- Description: Doctor of Philosophy
- Description: Structural optimisation is an important field of applied mathematics, which has proved useful in engineering projects. Reliability-based design optimisation (RBDO) can be considered a branch of structural optimisation. Different RBDO approaches have been applied in real world problems (e.g. vehicle side impact model, short column design, etc.). Double-loop, single-loop, and decoupled approaches are three categories in RBDO. This research focuses on double-loop approaches, which consider reliability analysis problems in their inner loops and design optimisation calculations in their outer loops. In recent decades, double-loop approaches have been studied and modified in order to improve their stability and efficiency, but many shortcomings still remain, particularly regarding reliability analysis methods. This thesis will concentrate on development of new reliability analysis methods that can be applied to solve RBDO problems. As a local optimisation algorithm, the conjugate gradient method will be adopted. Furthermore, a new method will be introduced to solve a reliability analysis problem in the polar space. The reliability analysis problem must be transformed into an unconstrained optimisation problem before solving in the polar space. Two methods will be introduced here and their stability and efficiency will be compared with the existing methods via numerical experiments. Next, we consider applications of RBDO models to electricity networks. Most of the current optimisation models of these networks are categorised as deterministic design optimisation models. A probabilistic constraint is introduced in this thesis for electricity networks. For this purpose, a performance function must be defined for a network in order to define safety and failure conditions. Then, new non-deterministic design optimisation models will be formulated for electricity networks by using the mentioned probabilistic constraint. These models are designed to keep failure probability of the network below a predetermined and accepted safety level.
Relationships between pumping costs and water quality in optimal operation of regional multiquality water distribution systems
- Authors: Mala-Jetmarova, Helena
- Date: 2016
- Type: Text , Thesis , PhD
- Full Text:
- Description: The operation of regional multiquality water distribution systems (WDSs) is a complex task involving multiple objectives in order to meet customer water quantity and quality requirements. These objectives, often conflicting, include scheduling of pumps to minimise pumping costs and mixing different quality waters from sources to ensure adequate quality water for customers. Evolutionary algorithms have been successfully applied to optimise operation of regional WDSs. Although a considerable reduction in pumping costs was demonstrated in past studies, other legitimate objectives, for example water quality, were not considered on an equal basis as they were included as a constraint. This single-objective approach precludes the tradeoffs between the objectives being obtained, so any insight on how to operate such a system cannot be provided should pumping costs and water quality be considered on equal basis. A multi-objective approach is applied in this thesis to optimise operation of regional multiquality WDSs considering pumping costs and water quality as legitimate objectives. Two optimisation models with increasing complexity are proposed. The first model considers two objectives, the pumping costs and a general water quality objective. The second model includes three objectives, the pumping costs and two water quality objectives for turbidity and salinity. The optimisation models are applied to three example networks from the literature using numerous scenarios and water quality data from the Wimmera Mallee Pipeline, Australia. A methodology is proposed to find the optimal solution for the multi‐objective optimisation of the WDS, which links a network simulator with a multi-objective genetic algorithm. Prior to optimisation, the performance of algorithm parameters is evaluated and their sensitivity analysed, for which a new methodology is developed. The following results were obtained. For the two-objective optimisation problem, there is a tradeoff with a competing nature between pumping costs and water quality. It means that reduction in pumping costs cannot be achieved without deterioration of water quality delivered to customers and vice versa. For the three-objective optimisation problem, interestingly, there is not a unique type of tradeoff (either competing or non-competing) between a particular pair of objectives. It is dependent on network hydraulics in combination with water quality at sources and customer water quality requirements. General principles behind the tradeoffs are formulated based on new categorisation of sources, so called consistent/inconsistent water quality (CWQ/IWQ) sources, in relation to customer water quality requirements. A practical approach for system operational strategy is developed for the purpose of long-term operational planning. It enables an operator to schedule supply from multiple sources with minimum pumping costs and customer water quality requirements being satisfied as much as possible, for all predicted water quality scenarios in the system.
- Description: Doctor of Philosophy
- Authors: Mala-Jetmarova, Helena
- Date: 2016
- Type: Text , Thesis , PhD
- Full Text:
- Description: The operation of regional multiquality water distribution systems (WDSs) is a complex task involving multiple objectives in order to meet customer water quantity and quality requirements. These objectives, often conflicting, include scheduling of pumps to minimise pumping costs and mixing different quality waters from sources to ensure adequate quality water for customers. Evolutionary algorithms have been successfully applied to optimise operation of regional WDSs. Although a considerable reduction in pumping costs was demonstrated in past studies, other legitimate objectives, for example water quality, were not considered on an equal basis as they were included as a constraint. This single-objective approach precludes the tradeoffs between the objectives being obtained, so any insight on how to operate such a system cannot be provided should pumping costs and water quality be considered on equal basis. A multi-objective approach is applied in this thesis to optimise operation of regional multiquality WDSs considering pumping costs and water quality as legitimate objectives. Two optimisation models with increasing complexity are proposed. The first model considers two objectives, the pumping costs and a general water quality objective. The second model includes three objectives, the pumping costs and two water quality objectives for turbidity and salinity. The optimisation models are applied to three example networks from the literature using numerous scenarios and water quality data from the Wimmera Mallee Pipeline, Australia. A methodology is proposed to find the optimal solution for the multi‐objective optimisation of the WDS, which links a network simulator with a multi-objective genetic algorithm. Prior to optimisation, the performance of algorithm parameters is evaluated and their sensitivity analysed, for which a new methodology is developed. The following results were obtained. For the two-objective optimisation problem, there is a tradeoff with a competing nature between pumping costs and water quality. It means that reduction in pumping costs cannot be achieved without deterioration of water quality delivered to customers and vice versa. For the three-objective optimisation problem, interestingly, there is not a unique type of tradeoff (either competing or non-competing) between a particular pair of objectives. It is dependent on network hydraulics in combination with water quality at sources and customer water quality requirements. General principles behind the tradeoffs are formulated based on new categorisation of sources, so called consistent/inconsistent water quality (CWQ/IWQ) sources, in relation to customer water quality requirements. A practical approach for system operational strategy is developed for the purpose of long-term operational planning. It enables an operator to schedule supply from multiple sources with minimum pumping costs and customer water quality requirements being satisfied as much as possible, for all predicted water quality scenarios in the system.
- Description: Doctor of Philosophy
Regulatory network discovery using heuristics
- Authors: Zarnegar, Armita
- Date: 2011
- Type: Text , Thesis , PhD
- Full Text:
- Description: This thesis improves the GRN discovery process by integrating heuristic information via a co-regulation function, a post-processing procedure, and a Hub Network algorithm to build the backbone of the network.
- Description: Doctor of Philosophy
- Authors: Zarnegar, Armita
- Date: 2011
- Type: Text , Thesis , PhD
- Full Text:
- Description: This thesis improves the GRN discovery process by integrating heuristic information via a co-regulation function, a post-processing procedure, and a Hub Network algorithm to build the backbone of the network.
- Description: Doctor of Philosophy