Dissertations / Theses on the topic 'Behaviour network'

Optical fibre communications are now widely used in many applications, including local area computer networks. I postulate that many future optical LANs will be required to operate with limited optical power budgets for a variety of reasons, including increased system complexity and link speed, low cost components and minimal increases in transmit power. Some developers will wish to run links with reduced power budget margins, and the received data in these systems will be more susceptible to errors than has been the case previously. The errors observed in optical systems are investigated using the particular case of Gigabit Ethernet on fibre as an example. Gigabit Ethernet is one of three popular optical local area interconnects which use 8B/10B line coding, along with Fibre Channel and Infiniband, and is widely deployed. This line encoding is also used by packet switched optical LANs currently under development. A probabilistic analysis follows the effects of a single channel error in a frame, through the line coding scheme and the MAC layer frame error detection mechanisms. Empirical data is used to enhance this original analysis, making it directly relevant to deployed systems. Experiments using Gigabit Ethernet on fibre with reduced power levels at the receiver to simulate the effect of limited power margins are described. It is found that channel bit error rate and packet loss rate have only a weakly deterministic relationship, due to interactions between a number of non-uniform error characteristics at various network sub-layers. Some data payloads suffer from high bit error rates and low packet loss rates, compared to others with lower bit error rates and yet higher packet losses. Experiments using real Internet traffic contribute to the development of a novel model linking packet loss, the payload damage rate, and channel bit error rate. The observed error behaviours at various points in the physical and data link layers are detailed. These include data-dependent channel errors; this error hot- spotting is in contrast to the failure modes observed in a copper-based system. It is also found that both multiple channel errors within a single code-group, and multiple error instances within a frame, occur more frequently than might be expected. The overall effects of these error characteristics on the ability of cyclic redundancy checks (CRCs) to detect errors, and on the performance of higher layers in the network, is considered. This dissertation contributes to the discussion of layer interactions, which may lead to un-foreseen performance issues at higher levels of the network stack, and extends it by considering the physical and data link layers for a common form of optical link. The increased risk of errors in future optical networks, and my findings for 8B/10B encoded optical links, demonstrate the need for a cross-layer understanding of error characteristics in such systems. The development of these new networks should take error performance into account in light of the particular requirements of the application in question.

The mechanisms and functions of shark grouping behaviour have received relatively little scientific attention to date. The current widespread use of social network analysis to study animal groups, in concert with rapid advances in animal tracking technology, now allows us to test specific hypotheses about how and why sharks form groups. This thesis uses replicated laboratory experiments to investigate some of the mechanisms underpinning aggregation in a model species of benthic, oviporous elasmobranch, the small spotted catshark (Scyliorhinus canicula L. 1758; Scyliorhinidae). Acoustic tracking of this species in the wild is also conducted to explore how network analyses can be adopted to study the localized movements, habitat connectivity and ranging behaviour of adult sharks. Groups of juvenile S. canicula were characterized by non-random social preferences, crucially, only when individuals were familiar with one another suggesting social recognition is important in young sharks of this species. Genetic analyses of parent and offspring DNA revealed very high levels of multiple paternity in this species, likely due to male sexual harassment and multiple mating, which leads to increased genetic diversity between juvenile sharks. Perhaps unsurprisingly, there was no evidence of kin relatedness structuring social interactions between conspecifics. Furthermore, testing the effects of environment on social behaviour provided evidence that these juveniles aggregated more in structurally complex environments than simple ones. However, at the individual level sharks showed consistency in their social network positions through time and across different habitat types. This result is indicative of personality traits in S. canicula. Using data gathered via passive acoustic telemetry of wild shark behaviour, network analysis provided a useful tool with which to quantify movement between receivers. One chapter has been dedicated to the application of these methods, highlighting a number of different analyses for predicting movement behaviour from such data. Finally, these methods were adopted to address ecological questions in this sexually segregated species. Persistent site fidelity to a localised inshore area by both male and female sharks suggested that segregation occurred at a relatively small spatial and temporal scale. Despite strong evidence of segregation, analyses of movement networks and individual co-occurrences revealed distinct periods of behavioural synchronicity during the months of March, April and May. In addition, habitat complexity appeared to be a significant driver of female behavioural strategy. Enhancing our knowledge of the social and environmental drivers of aggregation and movement in sharks is of great importance given the ecological threat facing many of our ocean’s top elasmobranch predators.

A new method is developed to determine a set of informative and refined interface assertions satisfied by functions that are represented by feed-forward neural networks. Neural networks have often been criticized for their low degree of comprehensibility.It is difficult to have confidence in software components if they have no clear and valid interface description. Precise and understandable interface assertions for a neural network based software component are required for safety critical applications and for theintegration into larger software systems. The interface assertions we are considering are of the form "e if the input x of the neural network is in a region (alpha symbol) of the input space then the output f(x) of the neural network will be in the region (beta symbol) of the output space "e and vice versa. We are interested in computing refined interface assertions, which can be viewed as the computation of the strongest pre- and postconditions a feed-forward neural network fulfills. Unions ofpolyhedra (polyhedra are the generalization of convex polygons in higher dimensional spaces) are well suited for describing arbitrary regions of higher dimensional vector spaces. Additionally, polyhedra are closed under affine transformations. Given a feed-forward neural network, our method produces an annotated neural network, where each layer is annotated with a set of valid linear inequality predicates. The main challenges for the computation of these assertions is to compute the solution of a non-linear optimization problem and the projection of a polyhedron onto a lower-dimensional subspace.

Doctor of Philosophy (PhD)
The aim of this thesis is to understand what drives the evolution of industrial networks and how such understanding can be used to stimulate sustainable development. A complex adaptive systems perspective has been adopted to analyse the complex interaction between organisational behaviour and industrial network evolution. This analysis has formed the basis for the development of a modelling approach that allows for quantitative exploration of how different organisational perceptions about current and future uncertainty affect their behaviour and therefore the network evolution. This analysis results in a set of potential evolutionary pathways for an industrial network and their associated performance in terms of sustainable development. Subsequently, this modelling approach has been used to explore the consequences of interventions in the network evolution and to identify robust interventions for stimulating sustainable development of industrial networks. The analysis, modelling approach and development of interventions has been developed in the context of a bioenergy network in the region of KwaZulu-Natal in South Africa. Industrial networks are an important aspect of today’s life and provide many goods and services to households and individuals all over the world. They consist of a large number of autonomous organisations, where some organisations contribute by transforming or transacting natural resources, such as oil, agricultural products or water, while other organisations contribute to networks by providing information or setting regulation or subsidies (local or national governments) or by influencing decision making processes of other organisations in networks (advocacy groups). Throughout the process from natural resource to product or service, industrial networks have important economic, environmental and social impacts on the socio-economic and biophysical systems in which they operate. The sum of complex interactions between organisations affects the rate in which natural resources are used, environmental impacts associated with transformation and transaction of resources and social impacts on local communities, regions or countries as a whole. The aim of this thesis is to understand how industrial networks evolve and how they can be stimulated towards sustainable development. The first question that has been addressed in this thesis is how to understand the complex interaction between organisational behaviour and industrial network evolution. Organisational behaviour is affected by many functional and implicit characteristics within the environment in which the organisation operates, while simultaneously the environment is a function of non-linear relationships between individual organisational actions and their consequences for both the function and structure of the network. This thesis has identified four different characteristics of industrial networks that affect organisational behaviour: 1) Functional characteristics 2) Implicit behavioural characteristics 3) Implicit relational characteristics 4) Implicit network characteristics. Functional characteristics are those characteristics that are formally recognised by all organisations within an industrial network and which affect their position within the network. Examples of functional characteristics are the price and quantity of resources available, the location and distance of organisations within a network, infrastructure availability or regulation. Implicit characteristics, on the other hand, are those characteristics that impact the decision making process of organisations, but which are not formally part of the network. From an organisational perspective, implicit characteristics are the rules, heuristics, norms and values that an organisation uses to determine its objectives, position and potential actions. Implicit relational characteristics, most importantly trust and loyalty, affect an organisations choice between potential partners and implicit network characteristics are those social norms and values that emerge through social embeddedness. Collectively, these functional and implicit characteristics and their interactions determine the outcome of organisational decisions and therefore the direction of the industrial network evolution. The complex interaction between these large numbers of characteristics requires quantitative models to explore how different network characteristics and different interactions result in different network evolutions. This thesis has developed an agent-based simulation model to explore industrial network evolutions. To represent the multi-scale complexity of industrial networks, the model consists of four scales. Each scale represents different processes that connect the functional and implicit characteristics of an industrial network to each other. The two basic scales represent the strategic actions of the organisations on the one hand and the industrial network function and structure on the other. The third scale represents the processes that take place within the mental models of organisations describing how they make sense of their environment and inform their strategic decision making process. The fourth scale represents the social embeddedness of organisations and how social processes create and destroy social institutions. The model has been developed such that it allows for exploring how changes in different network characteristics or processes affect the evolution of the network as a whole. The second question that has been addressed in this thesis is how to evaluate sustainable development of different evolutionary pathways of industrial networks. First of all, a systems approach has been adopted to explore the consequences of an industrial network to the larger socio-economic and biophysical system in which the network operates. Subsequently, a set of structural indicators has been proposed to evaluate the dynamic performance of industrial networks. These four structural indicators reflect the efficiency, effectiveness, resilience and adaptiveness of industrial networks. Efficiency and effectiveness relate to the operational features by which industrial networks provides a particular contribution to society. Resilience and adaptiveness relate to the system’s capacity to maintain or adapt its contribution to society while under stress of temporary shocks or permanent shifts, respectively. Finally, different multi-criteria decision analysis (MCDA) tools have been applied to provide a holistic evaluation of sustainable development of industrial networks. The third important question that is addressed in this thesis is how to systematically explore the potential evolutionary pathways of an industrial network, which has led to the development of agent-based scenario analysis. Agent-based scenario analysis systematically explores how industrial network evolutions might evolve depending on the perceptions of organisations towards the inherent uncertainty associated with strategic decision making in networks. The agent-based scenario analysis consists of two steps. Firstly, analysts develop a set of coherent context scenarios, which represents their view on the context in which an industrial network will operate within the future. For a bioenergy network, for example, this step results in a set of scenarios that each represent a coherent future of the socio-economic system in which the network might evolve. The second step is the development of a set of ‘agent scenarios’. Each agent-based scenario is based on a different ‘mental model’ employed by organisations within the network about how to deal with the inherent ambiguity of the future. The organisational perspective towards uncertainty is of major importance for the evolution of industrial networks, because it determines the innovative behaviour of organisations, the structure of the network and the direction in which the network evolves. One the one hand, organisations can ignore future ambiguity and base their actions on the environment that they can observe in their present state. On the other extreme, organisations can adopt a view that the future is inherently uncertain and in which they view social norms and values more important than functional characteristics to make sense of their environment. The mental models are differentiated according to two dimensions: 1) different mental representation of the world and 2) different cognitive processes that can be employed to inform strategic actions. Along these dimensions, different processes can be employed to make sense of the environment and to inform decision making. The thesis has shown that by systematically exploring the different perceptions possible, an adequate understanding of the different evolutionary pathways can be gained to inform the evaluation and development of interventions to stimulate sustainable development. The final part of this thesis has applied the analysis and methodology developed throughout this thesis to a bioenergy network in the province of Kwazulu-Natal in South Africa. The bioenergy network consists of a set of existing sugar mills with large quantities of bagasse, a biomass waste product, available. Bagasse is currently burned inefficiently to produce steam for the sugar mills, but can potentially be used for the production of green electricity, biodiesel, bioethanol or gelfuel. All of these products have important consequences for the region in terms of associated reductions in CO2 emissions, electrification of and/or energy provision for rural households and local economic development of the region. This thesis has modelled strategic decisions of the sugar mills, the existing electricity generator, potential independent energy producers, local and national governments and how their actions and interactions can lead to different evolutionary pathways of the bioenergy network. The agent-based scenario analysis has been used to explore how different perceptions of organisations can lead to different network evolutions. Finally, the model has been used to explore the consequences of two categories of interventions on stimulating sustainable development. The conclusions are that both categories of interventions, financial interventions by national government and the introduction of multi-criteria decision analysis (MCDA) tools to aid strategic decision making, can have both positive and negative effects on the network evolutions, depending on what ‘mental models’ are employed by organisations. Furthermore, there is no single intervention that outperforms the others in terms of stimulating both functional and structural features of sustainable development. The final conclusion is that instead of focusing on individual or collective targets, emphasis should be placed on the development of interventions that focus on evolutionary aspects of industrial networks rather than functional performance criteria. This thesis has also highlighted interesting research questions for future investigation. The methodology developed in this thesis is applied to a single case study, but there are still many questions concerning how different industrial networks might benefit from different organisational perceptions towards uncertainty. Furthermore, the role between the mental models and sustainable development requires further investigation, especially in the light of globalisation and the interconnectiveness of industrial networks in different countries and continents. Finally, this methodology has provided a platform for investigating how new technologies might be developed that anticipate needs of future generations. This thesis has provided a first and important step in developing a methodology that addresses the complex issues associated with sustainable development, benefiting both academics and practitioners that aim to stimulate sustainable development.

Thesis (Ph. D.)--University of Sydney, 2008.
Includes graphs and tables. Title from title screen (viewed December 17, 2008). Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy to the School of Chemical and Biomolecular Engineering, Faculty of Engineering and Information Technologies. Includes bibliographical references. Also available in print form.

In the last years the number of devices connected to the Internet has been increasing exponentially and it has reached huge numbers, just days ago the RIPE Network Coordination Center announced it run out of IPv4 addresses. As a consequence, it's extremely complex to control what is happening on the network and what, or whom, a device is communicating with. We are losing control over many of our devices. It's becoming everyday more difficult to know what the device is doing and what is it sharing on the Internet. Luckily, in the last years the attention to security, privacy and awareness is ever increasing: the users pay more attention to what their devices are doing and care about it. While big software (like the operative systems and most mobile devices) make available an increasing number of tools to monitor the device traffic, some devices cannot be inspected or those same tools may be forged in a way that makes it impossible to detect some specific malicious behaviours. Until now every successful method to detect or filter the behaviour of the devices has been trying to only detect anomalies or manually lock some specific behaviours; this means nobody ever tried to understand at a discrete level of accuracy what the device does relying only on the analysis of the metadata of the traffic intercepted. That’s why in this thesis I defined a model able to detect the behaviour which is occurring on the device by the mere observation of its network traffic. In doing so, I first defined a model that takes the raw low-level information regarding the communications occurring, process them and return information about the high-level operations occurring on the device. I then built a demo that uses that model and demonstrates its feasibility. The data used are low-level information on the traffic without ever inspecting its payload. To improve the efficiency of the algorithm, I also use a form of Recurrent Neural Networks to finally infer the device behaviour.

Individuals engage in competitive and cooperative interactions with conspecifics. Furthermore, within any population of interacting individuals there are typically consistent differences among-individuals in behavioural traits. Understanding the importance of both these types of individual-specific behaviours allows us to understand why populations are structured as they are, why individuals show apparently limited behavioural flexibility, and how these elements link to population-level properties. I used extensive video camera monitoring of a population of wild field crickets (Gryllus campestris) to study the interactions and behaviours of uniquely identified individuals. I studied the shyness, activity and exploration of individuals of this population across contexts: from young to old and between captivity and the wild. This allowed me to confirm that individuals were relatively consistent across their adult lifetimes for all three traits, but only consistent between captivity and the wild for activity and exploration. I then found that high activity levels were positively related to high mating rates and short lifespans. Crucially, lifetime mating success was not related to activity level, indicating that the trade-off between lifespan and mating success was sufficient to allow variation in activity level to persist across generations. I also found that cricket social network structure is stable across generations despite the complete turnover of individuals every year. This social network structure influences sexual selection, with some male crickets heavily involved in networks of both pre- and post-copulatory competition, yet males are unable to use pre-copulatory competition to avoid post-copulatory competition. Additionally, positive assortment by mating rate between males and females may reduce the fitness of males with high mating rates, as they face stronger sperm competition. Finally, I used actor-based models to determine the factors predicting cricket social network structure and to test and reject the social-niche hypothesis for the maintenance of among-individual variation in behaviour. I also demonstrated that little else is needed in a stochastically changing network aside from positive assortment by mating rate to simulate a population with a similar skew in mating success to the one observed in the real cricket population. These results give insights into the importance of trade-offs and stochasticity in maintaining the extensive variation in the natural world.

The development of fin damage in Atlantic salmon parr was investigated using social network analysis of behavioural interactions occurring under different feeding and stocking conditions. Four separate experiments were carried out in which groups of fish were subjected to a long food restriction period (30 days) described in Chapter 2, high (30 kg/m3) or low (8 kg/m3) fish stocking densities (Chapter 3), predictable or unpredictable food delivery regime (Chapter 4) or a short food restriction period (10 days) described in Chapter 5. Dorsal fin damage (erosion, splits and fin index) was significantly higher in groups of fish subjected to food restriction periods (short and long), held at high stocking density (30 kg/m3) or with an unpredictable food delivery regime. No other fins were found to be affected by fin damage irrespective of the treatment. The social networks based on aggressive interactions showed higher centrality, clustering coefficients, in-degree centrality, out-degree centrality and less dense networks in groups subjected to food restriction (short and long), low stocking densities 8 kg/m3), and unpredictable food delivery. The high centralities and clustering coefficients indicated separation of fish within the groups into initiators of aggression and receivers of aggression. This separation of roles was seen only in the food restricted group, high density groups and unpredictable food delivery groups. Initiators had higher out-degree centrality while receivers showed high in-degree centrality. Also, initiators of aggressive interactions had less fin erosion, higher final weights and higher body lengths than receivers of aggression. The severity of fin damage was significantly higher when this role differentiation occurred, and it was highly correlated with fin biting events. Additionally, overall aggression was higher in food restricted groups, low density groups and unpredictable groups. The dynamic analyses of networks over time (Chapter 6) showed that fish classified as initiators of aggression did not change this behaviour after normal/control environmental conditions were restored (i.e. ad libitum food delivery, low stocking density or predictable food delivery). The latter result indicates that individual fish maintained their behaviour irrespective of their social context, which is consistent with the definition of behavioural syndromes, personalities or temperament defined for other non-human animals. Overall these studies demonstrated the importance of using social network analysis to clearly identify and quantify roles that individual fish assume within their network group and through time based on their behavioural interactions leading to dorsal fin damage and differential physical characteristics. The results are potentially applicable within the commercial aquaculture industry as a valuable technique to evaluate and improve the welfare of farmed fish.

With the increased global use of online media platforms, there are more opportunities than ever to misuse those platforms or perpetrate fraud. One such fraud is within the music industry, where perpetrators create automated programs, streaming songs to generate revenue or increase popularity of an artist. With growing annual revenue of the digital music industry, there are significant financial incentives for perpetrators with fraud in mind. The focus of the study is extracting user behavioral patterns and utilising them to train and compare multiple supervised classification method to detect fraud.  The machine learning algorithms examined are Logistic Regression, Support Vector Machines, Random Forest and Artificial Neural Networks. The study compares performance of these algorithms trained on imbalanced datasets carrying different fractions of fraud. The trained models are evaluated using the Precision Recall Area Under the Curve (PR AUC) and a F1-score. Results show that the algorithms achieve similar performance when trained on balanced and imbalanced datasets. It also shows that Random Forest outperforms the other methods for all datasets tested in this experiment.
Med den ökande användningen av strömmande media ökar också möjligheterna till missbruk av dessa platformar samt bedrägeri. Ett typiskt fall av bedrägeri är att använda automatiserade program för att strömma media, och därigenom generera intäkter samt att öka en artist popularitet. Med den växande ekonomin kring strömmande media växer också incitamentet till bedrägeriförsök. Denna studies fokus är att finna användarmönster och använda denna kunskap för att träna modeller som kan upptäcka bedrägeriförsök. The maskininlärningsalgoritmer som undersökts är Logistic Regression, Support Vector Machines, Random Forest och Artificiella Neurala Nätverk. Denna studie jämför effektiviteten och precisionen av dessa algoritmer, som tränats på obalanserad data som innehåller olika procentandelar av bedrägeriförsök. Modellerna som genererats av de olika algoritmerna har sedan utvärderas med hjälp av Precision Recall Area Under the Curve (PR AUC) och F1-score. Resultaten av studien visar på liknande prestanda mellan modellerna som genererats av de utvärderade algoritmerna. Detta gäller både när de tränats på balanserad såväl som obalanserad data. Resultaten visar också att Random Forestbaserade modeller genererar bättre resultat för alla dataset som testats i detta experiment.

The concept of communities and the interaction between people are not new concepts. People have always gathered around common conditions shared by those in the group such as shared emotions, interests, beliefs and needs. It is however the way we interact, with whom we interact, and when and where these meetings take place which has changed. This has been a direct result of the development of the internet and exacerbated with the move to the second phase of internet development. This second phase of internet development provides users with real-time functionality enabling interaction with global users in a virtual environment. This interaction is termed online social networking and takes place in online communities.Online communities present opportunities for marketers as they give rise to a virtually unlimited number of different consumers, structured around finer consumption and marketing interests. Research indicates that online community users are market-oriented and therefore online communities provide a meaningful medium of exchange for these users. The challenge for companies today has been trying to develop ways to capitalise on this trend and raise their competitive advantage. However, in order to effectively understand these users, an understanding of their characteristics is fundamental to the development of any tailored marketing campaign. This thesis therefore aims to shed an insight into a segmentation model designed for online communities - firstly by empirically testing it and secondly, by enriching the data with a typology of online buying behaviour characterised by psychographic and behaviour variables. Therefore the research question posed is “Can online community users be classified by their online buying behaviours so that they are useful to marketers?”.The empirical data was gathered quantitatively through an online questionnaire designed to classify the respondents into meaningful segments and clusters. The report reflects a social constructionist methodology where the results have been interpreted and given meaning. The report is based on the segmentation models presented by Kozinets’ ‘virtual communities of consumption’ and Barnes et al. typology on online buying behaviour. This report combines the two models in order to enrich the segmentation model presented by Kozinets’ with attributes of online buying behaviour in order to provide a more comprehensive understanding of online community users. The results indicated that the four online user profiles defined in Kozinets’ model did not show differences in their online purchasing behaviour. Rather all online community users could be categorised by Barnes et al’s three clusters of online buying behaviour.

Port-knocking is a security mechanism used in computer systems to hide available network services. Its operation relies on a drop policy firewall setting in order to make impossible for port-scanning attacks to occur. This project researches the impact of implementing such a software solution. Furthermore, it looks into the behavior of three chosen implementations and make conclusions on the benefits and disadvantages that they bring. In addition, the surrounding implications related to both user and administrator are explored. This thesis includes tests on the resource consumption of the implementations as well as records of the added delay of using the mechanism when initiating a SSH session. There has not been such research performed in this field and the results of it could be beneficial to those who are involved in computer science and network security in particular. Finally, the product of this study state that port-knocking is overlooked and has great benefits in preventing zero-day exploits and hacker tools relying on exposed network services.

Innovation is defined as the introduction of new things or methods. In the history of human society, progress and cultural evolution occurred as a consequence of innovation processes. Typically changes proposed by a restricted number of peoples became widely adopted innovations as soon as a broad consensus formed around their adoption. In this thesis, we explore the role of innovators as potentially influential individuals in our society. Studying their behaviour is crucial to understand what are the factors that drove their decision in the process of becoming influential. In particular, here we uncover the importance of cultural attractors as cities where strong akin communities are present. Our approach involves the use of Wikipedia as a source for historical mobility data to model the migration patterns of globally relevant innovators. While here we study mobility on a broad range of different disciplines where different innovators gave their contributions, we also focus on a smaller and more modern system. Historical innovators are easily identified and discerned from uninfluential ones thanks to the wisdom of the time. However, due to the scarce availability of individual historical data, we point our attention to more recent versions of innovators: code developer. The flourishment of the digital era made code developers at the very centre of our global economy. We study this coupled system as a representative example of the interaction between innovators and the economy. Indeed, a significant, non-trivial interaction is found among the two worlds. More in general, in this thesis we highlight the relevance of innovators in shaping human collective responses. Our results reveal that innovators play a major role both individually and collectively at different scales. We provide measures of these effects (i) by looking at how innovator communities construct the attractiveness of a city and (ii) by studying how individual contributions in the innovation domain can dramatically affect financial behaviour also at short time scales. Our result expands the evidence of the need for a new research dimension, where human behaviour is studied as a complex system moving over an intricate network of intertwined interactions.

This research investigated the association of street network connectivity differences across travel modes with travel behaviour – mode choice, distance traveled and number of trips. To date research on travel behaviour relationships with urban form has not developed empirical evidence on street designs as distinct networks for walking and driving. A street network having greater connectivity and continuity for the pedestrian mode of travel vis-à-vis the vehicular network, like the Fused Grid, will likely encourage more walking. This hypothesis was investigated using a quasi-experimental approach within a rational utility behavioural framework. Local travel behaviour is theorized to be affected by desire to access goods and services (broadly termed, ‘activities’) in the community where people live. Using inferential statistics, the research tested for relationships between measured street patterns and self-reported local travel by King County, WA households. The main variables were ratios (walking : driving) of network connectivity and density, in the vicinity of travel survey households. Demographics and household characteristics, as well as other behaviourally influential urban form factors (residential density, proximity of destinations, etc.), were included in regression models, allowing control for confounding factors. Findings suggest that street networks with connectivity that provides better routing for one mode of transportation over others encourage more travel by the favored mode. The regression model demonstrated that a change from a pure small-block grid to a modified grid (i.e. Fused Grid) can result in an 11.3% increase in odds of a home-based trip being walked. The modified street pattern like a Fused Grid is also associated with a 25.9% increase, over street patterns with equivalent route directness for walking and driving, in the odds a person will meet recommended levels of physical activity. Finally, the Fused Grid’s 10% increase in relative connectivity for pedestrians is associated with a 23% decrease in local vehicle travel distance (VMT), and its improved continuity is associated with increased walking trips and distance. Conclusions: Other factors being equal, residential street networks with either more direct routing for pedestrians or more pedestrian facilities relative to vehicular network are associated with improved odds of walking and reduced odds of driving.

The robust behavioural courtship ritual displayed by Drosophila melanogaster males is governed by their underlying nervous system (NS). Two key genes of the Sex Determination Hierarchy, fruitless (fru) and doublesex (dsx), determine most neuronal substrates for sexual behaviour. In this study we aim to better understand the role fru plays in determining these neural substrates, as a means of better understanding the relationships between brain, behaviour and genes, and thus how the development of neuronal networks shape innate and species-specific behaviours. Fru has two major functions: control of male sexual behaviour, and viability in both sexes. Alternative splicing of fru produces transcription factors driven by four promoters: P1 transcripts are sex-specifically spliced (only viable in the male), and P2-4 transcripts are crucial to both sexes survival. The resulting proteins contain a BTB protein-protein interaction domain at the N-terminus, and one of four C-terminal zinc-finger (ZnF) DNA binding domains. Male-specific proteins (FruM) contain an additional 101 amino acid N-terminal domain, and one of three alternative C2H2 ZnF domains (FruMA, FruMB, FruMC). These male-specific isoforms are expressed in the central nervous system (CNS) beginning in the late stage larvae (L3), peaking during pupation and on into adulthood. Little is known, however, about the roles of the individual isoforms, and no clear transcriptional targets have been identified. The central aims of this thesis are to document the wild-type expression patterns of Fru isoforms throughout development in the CNS, create and characterise isoform-specific mutants, and to identify and evaluate putative transcriptional targets of Fru. Combining these findings will lead to a better understanding of the underlying molecular functions of individual FruM isoforms, as a means to understanding their roles in sexual behaviour. Expression analysis of FruM isoforms throughout development in the NS is described. To further characterise the role of individual FruM isoforms, isoform-specific mutants in fruA and fruB exons were generated using site-specific homologous recombination (HR). These novel mutants were validated by PCR and Fru isoform-specific antibody stainings. Mutants were analysed in fruM- and fru-null genetic backgrounds, to distinguish the roles of sex-specific vs. common isoforms. These analyses included: fertility, viability and morphology. FruA was found to have a role in wing extension with possible repercussions for song production. FruB was found to be developmentally lethal, in addition to having defects in male courtship behaviour. To understand the role of fru in the NS, downstream transcriptional targets of FruM isoforms were identified. DNA adenine methyltransferase identification (DamID) was used to identify putative transcription targets of FruM isoforms. The Dam protein methylates DNA in Drosophila in a sequence-specific manner allowing targets of Fru to be isolated. Candidate genes were identified using computational analysis (including gene ontology, peak analysis and motif analysis) along with a biologically significant connection with fru. The relationship between fru and six candidate genes were characterised using RNAi. The results of these studies advance our knowledge of how FruM isoforms influence the development and physiology of the NS underlying male sexual behaviour in Drosophila.

The emergence of collective patterns from repeated local interactions between individuals is a common feature to most living systems, spanning a variety of scales from cells to animals and humans. Subjects of this thesis are two aspects of emergent complexity in living systems: collective movement and transport network formation. For collective movement, this thesis studies the role of movement-mediated information transfer in fish decision-making. The second project on collective movement takes inspiration from granular media and soft mode analysis and develops a new approach to describe the emergence of collective phenomena from physical interactions in extremely dense crowds. As regards transport networks, this thesis proposes a model of network growth to extract simple, biologically plausible rules that reproduce topological properties of empirical ant trail networks.  In the second project on transport networks, this thesis starts from the simple rule of “connecting each new node to the closest one”, that describes ants building behavior, to study how balancing local building costs and global maintenance costs influences the growth and topological properties of transport networks. These projects are addressed through a modeling approach and with the aim of identifying minimal sets of basic mechanisms that are most likely responsible of large-scale complex patterns. Mathematical models are always based on empirical observations and are, when possible, compared to experimental data.

When a species is housed in captivity there are facets of the managed environment that can impact on individual and population welfare. A key component of an individual’s environment is its social environment, which can have important implications for animal health and welfare. Highly social species may experience impoverished welfare if kept in captive groups that differ in size, structure and demographic to that experienced by free-living animals. Amongst the most common of social animals to be housed in captivity are the flamingos (Phoenicopteridae). The unique evolutionary biology of these birds means that key aspects of their behaviour depends upon group living. A positive relationship between breeding success and increasing flock size has previously been noted, but how flamingo flocks are structured socially remains mysterious. All six species of flamingo are currently found in captive collections. The three more generalist species, greater (Phoenicopterus roseus), Caribbean (P. ruber) and Chilean (P. chilensis) flamingos are found commonly in zoos and can, under the right conditions, breed well. The other flamingo species are much more specialised in their habitat and dietary requirements and have proved more challenging with regards to their captive management. These three species are the Andean (Phoenicoparrus andinus), James’ (P. jamesi) and lesser (Phoeniconaias minor) flamingos. All six species are included in the experimental work presented in this thesis. This research aimed to evaluate specific elements of flamingo group living using social network analysis (SNA). It also aimed to assess influences of the captive environment (i.e. enclosure style and visitor number) and climate on enclosure usage, time-activity budgets and behavioural diversity to provide, as far as possible, a complete insight into how to measure, assess and evaluate captive flamingo welfare. A synthesis of the relevance of SNA to zoo animal management, and a review of current literature to identify research needs that could evidence good flamingo husbandry form the basis of the first two chapters. These two overview chapters support the questions asked in the following data-based sections of the thesis. Flamingos were observed at WWT Slimbridge Wetland Centre from March 2012 to July 2016 with data on patterns of social associations being collected four times daily (depending upon weather and bird husbandry). Associations were defined as birds within one neck length of each other and, using photos, the affiliations of each bird in the group were recorded. For assessment of bond strength, network position and identification of preferred/avoided partners a Half-Weight Index was applied to these data. Permutation testing was applied to association matrices to determine the difference between the number of observed preferential bonds (and avoided bonds) and Mantel tests were used to compare matrix correlations to assess differences between seasons, species, years and enclosures (where appropriate). The SNA programmes Socprog, UCInet and Netdraw were used to analyse network data. These network data form the basis of three chapters and show that flamingos associate preferentially with non-random bonds occurring in all flocks observed. Influences of social bonds on courtship display were also examined, and temporal changes in association were considered across time, season and year. Finally, to see any influence of animal health on bond preferences, scores of foot condition (used to identify and evaluate the presence and severity of pododermatitis on an individual bird) were analysed alongside of network measures for three flocks of flamingos. To measure enclosure usage, each species’ exhibit was measured and zones accessible to the birds were defined. As exhibit use can be based on resource use (and these resources can form differently-sized areas within an exhibit) a modified Spread of Participation Index (SPI) was used to provide an outcome between 1 (one area or resource used more than others) and 0 (equal use of all resource zones). Time-activity budgets were calculated for all flocks over daytime, and for one flock (measured using remote camera traps) over night. These data are presented in two chapters and demonstrate that captive flamingos can change their activity patterns in a similar manner to that noted in wild birds. Interestingly, flamingos are very active during the night and this provides useful data for zoo personnel to consider when re-assessing husbandry and management plans for these most ubiquitous of zoo birds.

Research into the social behaviour of wild animals living in groups has demonstrated the importance of social structure dynamics and their consequences for an individual’s fitness. Many aspects of animal behaviour and ecology, including interactions with conspecifics, habitat use and willingness to take risks, can be a reflection of personality. One of the key concerns of captive animal husbandry is the social environment, as it is regularly modified and can shape the social behaviour of the animals in question in different ways. In this thesis I explore how meerkat, Suricata suricatta, social dynamics and individual positions may differ between wild and captive groups; I explore personality in the context of social networks and, lastly, I explore how physical and husbandry factors vary across enclosures and how this corresponds to a variation in the social structure of meerkats. Differences were found between the fifteen groups of captive meerkats when considering association networks based on foraging and resting. Some of these differences could be explained by intrinsic differences between the groups. An individual’s position within a network as described by their centrality and closeness measures could be predicted by their age and status, but rarely by their sex. I did not detect consistent patterns of non-random assortment amongst group members based on their sex, age or status. Groups of wild and captive meerkats differed in various aspects of their social network structure. Such differences may be due to individuals occupying different network positions and the difference in their number and strength of their connections to other individuals. This distinct way of interacting and associating could be a result of group specific attributes, such as group size, and/or the attributes of the donor and recipient, including sex, status or age. Critically, the differences may be explained by the dissimilar living environment that each encounters. The current results suggest that a meerkat social network in captive conditions can be less consistent than in their wild environment in the way they associate with one another, and in the manner they occupy particular positions in the network. Principal component analysis of the four personality traits revealed two personality dimensions, Friendliness and Aggressiveness, across the fifteen groups of meerkats. However, within a subset of my data (five groups), Friendliness was the only measure that robustly captured consistent individual differences across at least one year. A relationship was not found between attributes and personality dimensions due to age, status, and sex. Individuals with high Friendliness scores were more central in networks of foraging competitions. Aggressiveness did not explain an individual’s position in any form of interaction. There was no evidence that meerkats preferentially associated with or avoid others based on each of their personality scores. A relationship was found in the way animals associate with one another in the resting network based on the size and complexity of the enclosure and the type of shelter. Individuals were less likely to associate with others of the same sex or dominance status in enclosures that were larger or more complex. All the six external measures (the size and complexity of the enclosure, the type of barrier and day shelter, environmental enrichment frequency and human contact) influenced how individuals interacted with other group members within grooming, playing and dominance networks. In general, it seems to be that the key features to address in meerkat management in zoos are those of enclosure size and complexity (and perhaps provision of adequate shelters). Providing captive meerkats with more naturalistic and complex enclosures can help to preserve their natural social system.

This thesis argues that while simplifications are a necessary part of the modelling process, there is a lack of empirical research to identify which types of variability should be included in our models, and how they should be represented. This research aims to develop methodologies to undertake empirical analyses of variability on the road network, focusing specifically on traveller behaviour. This is particularly timely given the emergence of rich new data sources. Firstly, a method is proposed for examining predictable differences in daily link flow profiles by considering both magnitude and timing. Unlike previous methods, this approach can test for statistically significant differences whilst also comparing the shapes of the profiles, by applying Functional Linear Models to transportation data for the first time. Secondly, a flexible, data-driven method is proposed for classifying road users based on their trip frequency and spatial and temporal intrapersonal variability. Previous research has proposed methodologies for identifying public transport user classes based on repeated trip behaviour, but equivalent methods for data from the road network did not exist. As there was not an established data source to use, this research examines the feasibility of using Bluetooth data. Spatial variability is examined using Sequence Alignment which has not previously been applied to Bluetooth data from road networks, nor for spatial intrapersonal variability. The time of day variability analysis adapts a technique from smart card research so that all observations are classified into travel patterns and, therefore, systematic and random variability can be measured. These network- and traveller-focused analyses are then brought together using a framework which uses them concurrently and interactively to gain additional insights into traveller behaviour. For each of the methods proposed, an application to at least one year of real world data is presented.

Advances in psychiatric genetics have begun to reveal the complex biological underpinnings of psychiatric disorders. Rare but penetrant copy number variants offer particularly direct mechanistic clues. The deletion at 15q11.2(BP1-BP2) has a 13% penetrance for developmental delay, congenital malformation, autism or schizophrenia. Reduced dosage of CYFIP1, one of four genes within this deletion, has emerged as a likely contributor to cognitive dysfunction seen in 15q11.2(BP1- BP2) deletion patients. However, the route from CYFIP1 haploinsufficiency to impaired behaviour has not been fully mapped. While synaptic deficits have been identified in mice haploinsufficient for Cyfip1 (Cyfip1+/-), circuit-level phenotypes have not been investigated. Using multi-site chronic electrode implants I recorded local field potential data simultaneously from prefrontal cortex, hippocampus and nucleus accumbens in a novel Cyfip1+/- rat model during a behavioural task and during sleep. Cyfip1+/- rats show normal performance accuracy on a discrete-trial alternation T maze task, but require more trials to achieve criterion during training. Task- dependent hippocampal-prefrontal network coordination remains intact in well-trained Cyfip1+/- rats, although theta-gamma phase-amplitude coupling within dorsal hippocampus is reduced compared to WTs. While circadian patterns and sleep architecture appear normal, hippocampal non-REM ripples are diminished in Cyfip1+/- rats compared to WTs, and preliminary data from the related Fmr1 (Fragile X Mental retardation 1) knockout rat also show aberrant ripples. Disrupted interactions are seen in the cortico-hippocampal-accumbal network, most prominently during approach to sucrose reward locations. Altered N-methyl-D- aspartate receptor signalling is implicated, as Cyfip1+/- rats show an exaggerated response to acute ketamine injection in the form of an enhanced surge in high frequency oscillations in nucleus accumbens and prelimbic cortex. Overall, abnormal behaviour- and ketamine-dependent network dynamics in hippocampus, prefrontal cortex and nucleus accumbens of Cyfip1+/- rats are reminiscent of some features of neuropsychiatric disorders, and lend weight to causal roles for CYFIP1 haploinsufficiency in predisposing patients to cognitive dysfunction.

From the individual to species level, it is common for animals to have connections with one another. These connections can exist in a variety of forms; from the social relationships within an animal society, to hybridisation between species. The structure of these connections in animal systems can be depicted using networks, often revealing non-trivial structure which can be biologically informative. Understanding the factors which drive the structure of animal networks can help us understand the costs and benefits of forming and maintaining relationships. Multivariate modelling provides a means to evaluate the relative contributions of a set of explanatory factors to a response variable. However, conventional modelling approaches use statistical tests which are unsuitable for the dependencies inherent in network and relational data. A solution to this problem is to use specialised models developed in the social sciences, which have a long history in modelling human social networks. Taking predictive multivariate models from the social sciences and applying them to animal networks is attractive given that current analytical approaches are predominantly descriptive. However, these models were developed for human social networks, where participants can self-identify relationships. In contrast, relationships between animals have to be inferred through observations of associations or interactions, which can introduce sampling bias and uncertainty to the data. Without appropriate care, these issues could lead us to make incorrect or overconfident conclusions about our data. In this thesis, we use an established network model, the multiple regression quadratic assignment procedure (MRQAP), and propose approaches to facilitate the application of this model in animal network studies. Through demonstrating these approaches on three animal systems, we make new biological findings and highlight the importance of considering data-sampling issues when analysing networks. Additionally, our approaches have wider applications to animal network studies where relationships are inferred through observing dyadic interactions.

The works reported in this thesis focus upon synthesising neural controllers for anthropomorphic robots that are able to manipulate objects through an automatic design process based on artificial evolution. The use of Evolutionary Robotics makes it possible to reduce the characteristics and parameters specified by the designer to a minimum, and the robot’s skills evolve as it interacts with the environment. The primary objective of these experiments is to investigate whether neural controllers that are regulating the state of the motors on the basis of the current and previously experienced sensors (i.e. without relying on an inverse model) can enable the robots to solve such complex tasks. Another objective of these experiments is to investigate whether the Evolutionary Robotics approach can be successfully applied to scenarios that are significantly more complex than those to which it is typically applied (in terms of the complexity of the robot’s morphology, the size of the neural controller, and the complexity of the task). The obtained results indicate that skills such as reaching, grasping, and discriminating among objects can be accomplished without the need to learn precise inverse internal models of the arm/hand structure. This would also support the hypothesis that the human central nervous system (cns) does necessarily have internal models of the limbs (not excluding the fact that it might possess such models for other purposes), but can act by shifting the equilibrium points/cycles of the underlying musculoskeletal system. Consequently, the resulting controllers of such fundamental skills would be less complex. Thus, the learning of more complex behaviours will be easier to design because the underlying controller of the arm/hand structure is less complex. Moreover, the obtained results also show how evolved robots exploit sensory-motor coordination in order to accomplish their tasks.

L'objectif de cette thèse est de trouver des modèles de choix d'itinéraire qui évoluent au niveau du réseau, c'est-à-dire des modèles qui rapprochent les choix des voyageurs sur la diversité des situations rencontrées dans un réseau de transport. L'approche de cette thèse pour étudier le comportement des voyageurs dans les réseaux de transport passe par des expériences informatiques à grande échelle, pour lesquelles une plateforme nommée Mobility Decision Game (MDG) a été développée. Le MDG permet d'observer les choix des participants sur un ensemble diversifié de scénarios (paires OD et itinéraires) avec des conditions de circulation et des informations de temps de trajet variables. Dans cette thèse, les expériences se concentrent sur les choix d'itinéraire des trajets en voiture qui sont basés sur la carte de la ville de Lyon, France. Pour atteindre l'objectif de cette thèse, une méthodologie de recherche de couples OD représentatifs du réseau est tout d'abord proposée. Les paires OD représentatives sont utilisées dans les expériences de choix de route pour obtenir des modèles de choix qui se généralisent aux différentes configurations OD dans le réseau. Deuxièmement, les choix des participants aux expériences sont analysés du point de vue du comportement rationnel et borné, afin d'établir le principe qui décrit le mieux leurs choix. Enfin, les modèles de choix sont évalués en fonction de leur précision prédictive. Cette thèse fait partie d'un projet européen ERC intitulé MAGnUM: Approche de modélisation du trafic multi-échelle et multimodale pour la gestion durable de la mobilité urbaine
The objective of this thesis is to find route choice models that scale-up at network level, i.e., models that predict the choices of travellers over the diversity of situations found in a transport network. The approach in this thesis to investigate travellers' behaviour in transportation networks is through computer-based experiments at large scale, for which a platform named the Mobility Decision Game (MDG), has been developed. The MDG permits to observe the choices of the participants on a diverse set of scenarios (OD pairs and routes) with varying traffic conditions and travel time information. In this thesis, the experiments focus on the route choices of uni-modal car trips that are based on the map of the city of Lyon, France. To attain the objective of this thesis, firstly a methodology to find OD pairs that are representative of the network is proposed. The representative OD pairs are used in route choice experiments to obtain choice models that generalise to the various OD configurations in the network. Secondly, the choices of participants in the experiments are analysed from the rational and boundedly rational behaviour perspectives, in order to establish the principle that best describe their choices. Finally, the choice models are assessed in terms of their predictive accuracy. This thesis is part of a European ERC project entitled MAGnUM: Multiscale and Multimodal Traffic Modeling Approach for Sustainable Management of Urban Mobility

We test the performance of myopic and farsighted stability concepts in a network formation experiment with a stream of payoffs and relatively unstructured link formation process. A subtle treatment variation demonstrates clearly the power of myopic stability concepts in precisely identifying the set of the most stable networks. However, we also find support for the predictions of farsighted concepts of stability, especially those that assume players' pessimism about the eventual outcome of a deviation. This is the first study to demonstrate that there exist environments where farsighted stability concepts identify empirically stable networks that are not identified by myopic concepts. Thus, myopic stability concepts are not necessarily sufficient to predict all stable outcomes in empirical applications.

This research describes information behaviour in context: experiences of serendipity in research. The study contributes to the understanding of serendipity as a complex phenomenon by looking at process in serendipity through a relational, phenomenological and sociological lens. The research asks: what linked events of doing and happening do people recount when they talk about their experiences of serendipity? and, how do they make sense of the circumstances surrounding these events? The research investigates a sample of fty rst-person narratives of lived experiences of serendipity recounted in the Citation Classics online dataset. A mixed methods parallel conversion design operationalises the research: one strand of the study focuses on description of contextual data, the other, on descriptions of two di erent event structure models. To meet its descriptive aims, the research draws on multiple methods: narrative approaches, network analysis and statistical techniques, including network topology inference and motif detection. A descriptive pro le of process in serendipity, a portfolio, which collects the network drawings and data for the one hundred event structures modelled by the study, and a research credibility audit stand as the study's substantive outcomes. The research fi ndings make a four-fold contribution to serendipity theory: they provide new insight into experiences of process in serendipity; add concrete, precise detail to fuzzy, abstract processrelated serendipity constructs; highlight problems with existing theoretical assumptions; and present evidence for normality in serendipity. Methodologically, the research opens alternative avenues into serendipity's complexity and brings fresh perspectives to the practice of serendipity research.

In recent years the advancement in machine learning have solved more and more complex problems. But still these techniques are not commonly used in the industry. One problem is that many of the techniques are black boxes, it is hard to analyse them to make sure that their behaviour is safe. This property makes them unsuitable for safety critical systems. The goal of this thesis is to examine if the deep learning technique Deep Q-network could be used to create a behaviour tree that can solve the same problem. A behaviour tree is a tree representation of a flow structure that is used for representing behaviours, often used in video games or robotics. To solve the problem two simulators are used, one models a cart that shall balance a pole called cart pole, the other is a static world which needs to be navigated called grid world. Inspiration is taken from the learning from demonstration field to use the Deep Q-network as a teacher and then create a decision tree. During the creation of the decision tree two attributes are used for pruning; to look at the trees accuracy or performance. The thesis then compare three techniques, called Naive, BT Espresso, and BT Espresso Simplified. The techniques are used to transform the extracted decision tree into a behaviour tree. When it comes to the performance of the created behaviour trees they all manage to complete the simulator scenarios in the same, or close to, capacity as the trained Deep Q-network. The trees created from the performance pruned decision tree are generally smaller and less complex, but they have worse accuracy. For cart pole the trees created from the accuracy pruned tree has around 10 000 nodes but the performance pruned trees have around 10-20 nodes. The difference in grid world is smaller going from 35-45 nodes to 40-50 nodes. To get the smallest tree with the best performance then the performance pruned tree should be used with the BT Espresso Simplified algorithm. This thesis have shown that it is possible to use knowledge from a trained Deep Q-network model to create a Behaviour tree that can complete the same task.
Under de senaste åren har ett antal framsteg inom maskininlärning gjorts vilket har lett till att mer och mer komplexa problem har kunnat lösas. Dock är dessa tekniker ofta inte använda av industrin. Ett av problemen är att många av de bättre teknikerna beter sig som svarta lådor, det är väldigt svårt att analyser vad de kommer att göra. Denna egenskap gör att de inte är lämpliga att användas i säkerhetskritiska system. Målet med denna avhandling är att undersöka möjligheten att använda den djupa inlärningstekniken djupa q-nätverk kan användas för att skapa ett beteendeträd som är kapabelt att lösa samma problem. Ett beteendeträd är en flödesstruktur som används för att representera beteenden, ofta använt i dataspel eller för robotar. För att undersöka problemet så används två simulatorer, den ena modellerar en vagn som ska balansera en stav och kallas vagnstav (cart pole). Den andra simulatorn är en statisk värld där målet för agenten är att ta sig till en definierad målplats, vilken kallas rutvärld (grid world). För att lösa problemet tas inspiration från ett angränsande fält kallat inlärning från demonstration. Istället för att använda en mänsklig lärare ansätts det djupa q-nätverket som lärare och används för att skapa ett beslutsträd. Beslutsträdet är sedan reducerat genom att kolla på trädets träffsäkerhet eller hur mycket belöning trädet får. Tre tekniker jämförs för att transformera beslutsträdet till ett beteendeträd, teknikerna heter Naiv, BT Espresso och BT Espresso förenklad. Alla skapade beteendeträd lyckas klara av problemet i simulatorn de är skapade för. De hade liknande prestanda som det djupa q-nätverket. När beslutsträden var reducerat på belöning resulterade det i generellt mindre beteendeträd, dock så hade de inte full träffsäkerhet mot det djupa q-nätverket. För vagnstav simulatorn hade beteendeträden som skapats från träffsäkerhets beslutsträden runt 10 000 noder, mot belönings kapade träd som hade runt 10–20 noder. I rutvärlden var skillnaden mindre med 40–50 noder för träd skapade från träffsäkerhet reducerade beslutsträde och 35–45 noder för belöning reducerade beslutsträd. Denna avhandling har påvisat att det går att skapa beteende träd från en tränad djup q-nätverksmodell för ett scenario och om det minsta trädet som klarar scenariot är att önskat bör belönings reducerade beslutsträd användas med BT Espresso förenkling algoritmen.

The occurence of financial contagion can lead to hazardous results for financial institutions, financial markets as well as for the whole economy. Therefore it can have even serious economic effects on everybody´s life. That is why it is of great interest to deeper understand its characteristics. As classical finance theory seems not to give the best answers to this topic, the young academic field of behavioural finance can deliver new insights. The main purpose of this work is to provide an introduction mainly to professionals in portfolio and risk management and help them to tackle the problem of contagion at an early stage. Therefore not only aspects of behavioural finance are discussed, but the topic contagion is also brought into connection with network analyses and the current regulation process. Our paper can not answer all questions related to contagion, but it can help the reader to better understand its main aspects and enables him to delve deeper into this field.

The maintenance of both the psychological and physiological health of captive animals is a key priority of modern zoos. Recognising that characteristics of the captive environment have the potential to decrease animal welfare, methods for quantifying and assessing welfare have been developed as part of the process for improving animal welfare. Traditionally, observations of animal behaviour and quantifying time budgets in relation to those of the animals’ wild counterparts have been utilised to assess animal welfare. Hormonal assays have also been implemented to quantify the physiological stress response of animals in captivity and identify the extent of stress being experienced. Each of these methods focuses on a different indicator of animal welfare, is quantified in different ways and provides a different perspective on the welfare of the animals. Given the limited time and financial budgets available to zoos and animal carers, identifying the most appropriate method of welfare assessment would be advantageous in helping to secure the best possible health of captive animals and to maximise their value in captivity. This thesis implemented both behavioural observations and hormonal assays to identify the strengths and weaknesses of each methodology, and make recommendations for future research. The study involved a group of four Bornean orangutans (Pongo pygmaeus) housed at Twycross Zoo. Behavioural observations involved continuous group sampling and the development of an ethogram to record a comprehensive account of orangutan activity over the course of a 12 week enrichment programme. Simultaneous to these observations, faecal samples were collected from each orangutan and processed via Enzyme Immunoassay (EIA) to quantify levels of faecal glucocorticoid metabolites (fGCM) in each sample. While recognising the recent developments in ecological analytical methods, the capacity for extending network analysis beyond the application to social networks, and its use as a welfare assessment tool were explored. Behavioural and space-use networks were developed using data from a second study of the orangutans housed at Twycross Zoo. The flexibility of network analysis in visually representing different data types allowed for the intuitive representation of complex behavioural data. Further research investigated the use of network metrics in providing deeper insights into animal behaviour and space use patterns. In addition, bipartite networks were assessed for their potential to detect and show patterns in the relationships between two sets of behavioural data. Each of the methods used had a number of strengths and weaknesses, but importantly each contributed a different perspective in the assessment of behaviour patterns and welfare, suggesting that an integrated approach to behaviour studies utilising several methods would be ideal. Cost and logistic constraints make this unlikely in most cases. However, the thesis ends with a look to the future and the recognition that the current rapid development of technology for use in animal behaviour studies, coupled with equally rapid development of analytical techniques, may help to dramatically increase the amount of information gained from the average animal behaviour study in the future. Such improvements have never been more urgent, with the requirement for understanding animal behaviour in light of current extinction rates within the context of habitat destruction and climate change. It is hoped that this thesis will make a contribution to improving future animal behaviour and welfare studies by providing an assessment of both traditional methods of study as well as demonstrating the use and potential of new ways of applying network analysis within such studies.

Ecological network theory has developed from studies of static, binary trophic relationships to the analysis of quantitative, dynamic communities consisting of multiple link-types. Particularly, work has focused on the dynamic nature of ecological networks, which maintains stability in complex communities. However, there are few in situ network-level studies of the determinants of temporal and spatial variation in community structure. This thesis utilises data from a 10-year study of a host-parasitoid network and a collaborative study in an applied ecological setting to identify individual level factors important to network structure. The work aims towards an empirical, predictive framework linking adaptive foraging behaviour to ecological network structure. The results show that condition-dependent foraging behaviours structure host-parasitoid networks. The realised niches of the studied parasitoid species were generally biased towards larger host species and condition-dependent sex ratio allocation increased the likelihood that females would eclose from relatively larger hosts and males from relatively smaller hosts, which resulted in sex ratios deviating from Hamiltonian (50:50) predictions. Further, both of these aspects of behaviour are plastic, where parasitoid behaviour responded to environmental heterogeneity. Particularly, host preference behaviour conformed to an egg-/time-limitation framework, where the size dependency of the behaviour is greater when individuals have a greater likelihood of being egg-limited. Both the size-dependency and the plasticity of these behaviours differed significantly between secondary parasitoid species. This species identity effect interacted with landscape heterogeneity, which may explain some inter- and intra-specific variation in network structure. With respect to applied ecology, the results show that the benefits of natural vegetation for pest control are dependent upon the dispersal capabilities and the diet breadth of the pest and its natural enemies. The findings are evaluated towards a predictive framework for understanding the effects of future climate change on community structure and stability. We consider this framework in terms of applied ecology, particularly pest control ecosystem services provided by natural vegetation in an agricultural environment. The synergistic nature of the multiple determinants of network structure found in this thesis suggest that future studies should focus on the whole network, which is not necessarily the sum of its parts.

The Internet has been more and more intensively used in recent years. Although network infrastructure has been regularly upgraded, and the ability to manage heavy traffic greatly increased, especially on the core networks, congestion never ceases to appear, as the amount of traffic that flow on the Internet seems to be increasing at an even faster rate. Thus, congestion control mechanisms play a vital role in the functioning of the Internet. Active Queue Management (AQM) is a popular type of congestion control mechanism that is implemented on gateways (most notably routers), which can predict and avoid the congestion before it happens. When properly configured, AQMs can effectively reduce the congestion, and alleviate some of the problems such as global synchronisation and unfairness to bursty traffic. However, there are still many problems regarding AQMs. Most of the AQM schemes are quite sensitive to their parameters setting, and these parameters may be heavily dependent on the network traffic profile, which the administrator may not have intensive knowledge of, and is likely to change over time. When poorly configured, many AQMs perform no better than the basic drop-tail queue. There is currently no effective method to compare the performance of these AQM algorithms, caused by the parameter configuration problem. In this research, the aim is to propose a new analytical model, which mainly uses discrete-time queueing theory. A novel transient modification to the conventional equilibrium-based method is proposed, and it is utilised to further develop a dynamic interactive model of responsive traffic and bottleneck queues. Using step-by-step analysis, it represents the bursty traffic and oscillating queue length behaviour in practical network more accurately. It also provides an effective way of predicting the behaviour of a TCP-AQM system, allowing easier parameter optimisation for AQM schemes. Numerical solution using MATLAB and software simulation using NS-2 are used to extensively validate the proposed models, theories and conclusions.

This thesis presents a research model to investigate university-students’ acceptance and behaviour in relation to social-networking sites (SNS). In order to carry out this investigation, the research project was divided into two phases using qualitative and quantitative data based on a diverse sample of university students. Phase One used a think-aloud technique to explore the interaction experiences associated with students’ use of Facebook, a popular social-networking site. Twenty-six participants from Teesside University took part in the first study and six categories of experience (communication, gratification, inquisitiveness, evocation, interconnection, apprehension, and ambience) were identified. Subsequently conceptual similarities were found between all six categories of experience and six psychological human needs (relatedness, pleasure, popularity, security, competence and meaning). In Phase Two, a research model was constructed, based on existing literature on technology acceptance and the psychological needs identified in Phase One. Results from an online survey of 766 university students in the United Kingdom, who were also SNS users, provided evidence for the proposed model. The model explained and predicted students’ adoption of SNS, accounting for half of the variance in behavioural intention and almost a quarter of the variance in actual use behaviour. The results showed that students’ personal beliefs, social identity and psychological human needs influenced their decision to adopt SNS. Specifically, user-perceived usefulness, ease of use, enjoyment and credibility were found to iv be important factors in students’ adoption of SNS. The influence of social identity on students’ behavioural intention was also found to be mediated by perceived usefulness and perceived ease of use. Furthermore, the findings emphasise the importance of psychological human needs in students’ adoption of SNS. In particular, the need for relatedness was found to be a significant independent predictor of behavioural intention. Based on the results of this study, theoretical and practical implications are discussed.

This thesis is based on the PhD work of investigating the Friction Stir Welding process (FSW) with numerical and Artificial Neural Network (ANN) modelling methods. FSW was developed at TWI in 1991. As a relatively new technology it has great advantages in welding aluminium alloys which are difficult to weld with traditional welding processes. The aim of this thesis was the development of new modelling techniques to predict the thermal and deformation behaviour. To achieve this aim, a group of Gleeble experiments was conducted on 6082 and 7449 aluminium alloys, to investigate the material constitutive behaviour under high strainrate, near solidus conditions, which are similar to what the material experiences during the FSW process. By numerically processing the experimental data, new material constitutive constants were found for both alloys and used for the subsequent FSW modelling work. Importantly no significant softening was observed prior to the solidus temperature. One of the main problems with numerical modelling is determining the values of adjustable parameters in the model. Two common adjustable parameters are the heat input and the coefficients that describe the heat loss to the backing bar. To predict these coefficients more efficiently a hybrid model was created which involved linking a conventional numerical model to an ANN model. The ANN was trained using data from the numerical model. Then thermal profiles were abstracted (summarised) and used as inputs; and the adjustable parameters were used as outputs. The trained ANN could then use abstracted thermal profiles from welding experiments to predict the adjustable parameters in the model. The first stage involved developing a simplified FE thermal model which represents a typical welding process. It was used to find the coefficients that describe the heat loss to the backing bar, and the amount of power applied in the model. Five different thermal boundary conditions were studied, including both convective and ones that included the backing bar with a contact gap conductance. Three approaches for abstracting the thermal curves and using as inputs to the ANN were compared. In the study, the characteristics of the ANN model, such as the ANN topology and gradient descent method, were evaluated for each boundary condition for understanding of their influences to the prediction. The outcomes of the study showed that the hybrid model technique was able to determine the adjustable parameters in the model effectively, although the accuracy depended on several factors. One of the most significant effects was the complexity of the boundary condition. While a single factor boundary condition (e.g. constant convective heat loss) could be predicted easily, the boundary condition with two factors proved more difficult. The method for inputting the data into the ANN had a significant effect on the hybrid model performance. A small number of inputs could be used for the single factor boundary condition, while two factors boundary conditions needed more inputs. The influences from the characteristics of the ANN model were smaller, but again thermal model with simpler boundary condition required a less complex ANN model to achieve an accurate prediction, while models with more complex boundary conditions would need a more sophisticated ANN model. In the next chapter, the hybrid method was applied to a FSW process model developed for the Flexi-stir FSW machine. This machine has been used to analyse the complex phase changes that occur during FSW with synchrotron radiation. This unique machine had a complex backing bar system involving heat transfer from the aluminium alloy workpiece to the copper and steel backing bars. A temperature dependent contact gap conductance which also depends on the material interface type was used. During the investigation, the ANN model topologies (i.e. GFF and MFF) were studied to find the most effective one. Different abstracting methods for the thermal curves were also compared to explore which factors (e.g. the peak temperature in the curve, cooling slope of a curve) were more important to be used as an input. According to close matching between the simulation and experimental thermal profiles, the hybrid model can predict both the power and thermal boundary condition between the workpiece and backing bar. The hybrid model was applied to six different travel speeds, hence six sets of heat input and boundary condition factors were found. A universal set was calculated from the six outcomes and a link was discovered between the accuracy of the temperature predictions and the plunge depth for the welds. Finally a model with a slip contact condition between the tool and workpiece was used to investigate how the material flow behaviour was affected by the slip boundary condition. This work involved aluminium alloys 6082-T6 and 7449-T7, which have very different mechanical properties. The application of slip boundary condition was found to significantly reduce the strain-rate, compared to a stick condition. The slip condition was applied to the Flexi-stir FSW experiments, and the results indicated that a larger deformation region may form with the slip boundary condition. The thesis successfully demonstrates a new methodology for determining the adjustable parameters in a process model; improved understanding of the effect of slip boundary conditions on the flow behaviour during FSW and insight in to the behaviour of aluminium alloys at temperatures approaching the solidus and high strain-rates.

Today the majority of the driver assistance systems are rule-basedcontrol systems that help the driver control the truck. But driversare looking for something more personal and exible that can controlthe truck in a human way with their own preferences. Machine learningand articial intelligence can help achieve this aim. In this studyArticial Neural Networks are used to model the driver steering behaviourin the Scania Lane Keeping Assist. Based on this, trajectoryplanning and steering wheel torque response are modelled to t thedriver preference. A model predictive controller can be used to maintainstate limitations and to weigh the two modelled driver preferencestogether. Due to the diculties in obtaining an internal plant modelfor the model predictive controller a variant of a PI-controller is addedfor integral action instead. The articial neural network also containsan online learning feature to further customize the t to the driverpreference over time.
Idag används till största del regelbaserade reglersystem förförarassistanssystem i lastbilar. Men lastbilschaufförer vill ha någotmer personligt och flexibelt, som kan styra lastbilen på ett mänskligtsätt med förarens egna preferenser. Maskininlärning och artificiell intelligenskan hjälpa till för att uppnå detta mål. I denna studie användsartificiella neurala nätverk för att modellera förarens styrbeteende genomScania Lane Keeping Assist. Med användning av detta modellerasförarens preferenser med avseende på placering på vägbanan och momentpåslag på ratten. En modell prediktiv kontroller kan användas föratt begränsa tillstånd och för att väga de två modellerade preferensernamot varann. Eftersom det var mycket svårt att ta fram den internaprocessmodellen som krävdes för regulatorn används istället en variantav en PI-kontroller för att styra lastbilen. De artificiella neuralanätverken kan också tillåtas att lära sig under körning för att anpassasig till förarens preferenser över tid.

L’eau est une ressource vitale, indispensable à la vie sur terre. A l’instar de nombreuses autres ressources naturelles, l’eau propre à la consommation est soumise à une forte pression à cause de l’impact de l’activité humaine d’une part et de l’augmentation continue de la population mondiale d’autre part. Une pression tellement forte que l’eau propre représente l’un des 17 objectifs de développement durable des Nations Unies. Dans ce contexte, une gestion rationnelle et durable de la ressource s’avère indispensable. Dans ce but, un système intelligent de supervision des réseaux d’eau potable peut s’avérer très utile. Les systèmes existant sont toutefois peu intégrés et compacts, nécessitent souvent une alimentation externe, et restent relativement chers pour un déploiement massif sur les réseaux. La présente thèse s’inscrit dans le cadre d’un projet de recherche européen, PROTEUS, visant à pallier ces différents problèmes en mettant au point un système de mesure pour la supervision de la ressource en eau permettant la mesure de 9 paramètres physico-chimiques, reconfigurable, et énergétiquement autonome. La contribution de la présente thèse à ce projet porte sur la conception et l’optimisation des différents capteurs physiques (conductivité électrique, pression, température et débit) ainsi qu’à leur co-intégration sur une même puce. Le système proposé montre des performances au moins égales à celle de l’état de l’art en ce qui concerne la robustesse, assurée par la redondance de nombreux éléments sensibles, le domaine de sensibilité et la consommation énergétique. Le présent manuscrit est par conséquent construit comme suit : le premier chapitre est une introduction générale à la supervision de grandeurs environnementales et à la puce multi-capteurs. Le second chapitre décrit la structure de la puce multi-capteurs ainsi que les méthodes de fabrication utilisées, avec une attention particulière accordée aux capteurs de pression et de conductivité électrique. Le troisième chapitre porte sur l’utilisation de résistances électriques pour la mesure de diverses grandeurs physiques, notamment la température. Le dernier chapitre s’attarde plus particulièrement sur l’utilisation de ce type de résistances pour la mesure de débit avant de conclure et de proposer des perspectives pour des travaux futurs
Water is a vital element for every living being on the earth. Like many other dwindling natural resources, clean water faces a strong pressure because of human activity and the rapid growth of global population. The situation is so critical that clean water has been identified as one of the seventeenth sustainable development goals of the United Nations. Under these conditions, a sustainable management of water resources is necessary. For this purpose, a smart solution for water networks monitoring can be very helpful. However, commercially available solutions lack compactness, self-powering capabilities cost competitiveness, necessary to enable the large rollout over water networks. The present thesis takes place in the framework of a European research project, PROTEUS, which addresses these different problems by designing and fabricating a multi-parameter sensor chip (MPSC) for water resources monitoring. The MPSC enables the measurement of 9 physical and chemical parameters, is reconfigurable and self-powered. The present thesis addresses more precisely physical sensors, their design, optimization and co-integration on the MPSC. The developed device exhibits state of the art or larger performances with regard to its redundancy, turn-down ratio and power consumption. The present manuscript is split into two main parts: Part-I and Part-II. Part-I deals with non-thermal aspects of the MPSC, the pressure and conductivity sensor for instance, as well as the fabrication process of the whole device (Chapter 1 and 2). The background of environmental monitoring is presented in Chapter 1 along with the State of Art review. Chapter 2 describes fabrication methods of the MPSC. Preliminary characterization results of non-thermal sensors are also reported in this chapter. Chapter 3 and 4, included in Part-II, deal with thermal sensors (temperature and flow-rate). Chapter 3 describes the many possible uses of electric resistances for sensing applications. Finally, in chapter four, we focus on flowrate sensors before concluding and making a few suggestions for future works

Modelling is an important technique in the comprehension and management of complex systems. Queueing network models capture most relevant information from computer system and network behaviour. The construction and resolution of these models is constrained by many factors. Approximations contain detail lost for exact solution and/or provide results at lower cost than simulation. Information at the resource and interactive command level is gathered with monitors under ULTRIX'. Validation studies indicate central processor service times are highly variable on the system. More pessimistic predictions assuming this variability are in part verified by observation. The utility of the Generalised Exponential (GE) as a distribution parameterised by mean and variance is explored. Small networks of GE service centres can be solved exactly using methods proposed for Generalised Stochastic Petri Nets. For two centre. systems of GE type a new technique simplifying the balance equations is developed. A very efficient "building bglloocbka"l. is presented for exactly solving two centre systems with service or transfer blocking, Bernoulli feedback and load dependent rate, multiple GE servers. In the tandem finite buffer algorithm the building block illustrates problems encountered modelling high variability in blocking networks. A parametric validation study is made of approximations for single class closed networks of First-Come-First-Served (FCFS) centres with general service times. The multiserver extension using the building block is validated. Finally the Maximum Entropy approximation is extended to FCFS centres with multiple chains and implemented with computationally efficient convolution.