Many approaches have been proposed for the recognition of mathematical formulae, traditionally using the results of optical character recognition over scanned documents. However, optical character recognition generally performs poorly when presented with mathematics, making it difficult to accurately parse formulae. Due to the rapidly increasing number of natively digital documents available, an alternative to optical character recognition is now available, that of analysing files directly instead of images. In this thesis, we explore such a method, analysing files in the ubiquitous Portable Document Format directly and combining it with image analysis, to produce the necessary information for the analysis of mathematical formulae and documents. We also revisit a method proposed in the 1960s for parsing handwritten mathematics. An extremely efficient, yet impractical approach due to a reliance of perfect input and precise character positioning. We heavily modify and extend this method, removing many of its restrictions and use it in conjunction with the perfect input from the PDF analysis, yielding high quality results which compare favourably with the leading scientific document analysis system.

Service oriented Architecture (SoA) is a layered architecture for organising software resources as services, so that they can be deployed, discovered and combined to produce new Web services. One of the key challenges of SoA is in identifying the occurrence of failures that may result in violations of Service Level Agreements, causing financial penalties or customer dissatisfaction. Therefore, it is crucial to develop methods of on-line detection of failure to take suitable remedial actions without delay. One of the methods of identifying occurrences of failure is to use Diagnosers; software modules which are deployed with the system to monitor the interaction between the services. This thesis presents a diagnostic approach for SoA based on extending the Diagnosability theory of Discrete Event System (DES). In particular, this research has resulted in a method of automated creation of Diagnosers and integrating them to the system. This is accomplished by coming up with an appropriate modelling language framework, which is a prerequisite to applying DES techniques. Modelling languages popular in DES, such as Petri nets and Automata, despite being sufficiently adequate for modelling, are not well adopted by the SoA community. Inspired by Petri nets and Workflow Graphs, a modeling approach, which closely follows BPEL, is proposed. Then, one of existing DES methods is extended for the creation of centralised Diagnoser. Various methods are proposed to implement and integrate the produced Diagnoser into the system. As a proof of concept, an implementation of the suggested approach is created as a Plugin for Oracle JDeveloper. A series of empirical results on the performance-related aspects of the proposed method are discussed.

This thesis presents a Cloud-based-Multi-Agent System (Clobmas) that uses multiple double auctions, to enable applications to self-adapt, based on their QoS requirements and budgetary constraints. We design a marketplace that allows applications to select services, in a decentralized manner. We marry the marketplace with a decentralized service evaluation and- selection mechanism, and a price adjustment technique to allow for QoS constraint satisfaction. Applications in the cloud using the Software-As-A-Service paradigm will soon be commonplace. In this context, long-lived applications will need to adapt their QoS, based on various parameters. Current service-selection mechanisms fall short on the dimensions that service based applications vary on. Clobmas is shown to be an effective mechanism, to allow both applications (service consumers) and clouds (service providers) to self-adapt to dynamically changing QoS requirements. Furthermore, we identify the various axes on which service applications vary, and the median values on those axes. We measure Clobmas on all of these axes, and then stress-test it to show that it meets all of our goals for scalability.

This is a study exploring the effects of blog assisted language learning (BALL) in the EFL writing classroom in Taiwan. It focuses principally on a comparative experimental project that was carried out over the period of one academic year. The project involved two groups of first year university students of a low intermediate level in English. There were 25 students in each group, and both groups were taught by the same teacher-researcher (i.e. the author of this thesis). In this thesis, we present a detailed statistical examination of various aspects of BALL, including students’ writing performance, learning motivation towards writing and self-efficacy beliefs with regard to their EFL writing skills and abilities. We also take a corpus-based look into the written language collected from both of the groups using a series of numerical counts obtained by computerised measurements. We also present the results of a qualitative, phenomenological analysis that reflects the nature of the participants’ BALL experience. The overall argument of the thesis is that BALL is at best no more effective than traditional approaches to EFL writing pedagogy, and in some cases is actually less effective than traditional, low-tech methods.

Worldwide, the number of people living with self-limiting conditions, such as Dementia, Parkinson’s disease and depression, is increasing. The resulting strain on healthcare resources means that providing 24-hour monitoring for patients is a challenge. As this problem escalates, caring for an ageing population will become more demanding over the next decade, and the need for new, innovative and cost effective home monitoring technologies are now urgently required. The research presented in this thesis directly proposes an alternative and cost effective method for supporting independent living that offers enhancements for Early Intervention Practices (EIP). In the UK, a national roll out of smart meters is underway. Energy suppliers will install and configure over 50 million smart meters by 2020. The UK is not alone in this effort. In other countries such as Italy and the USA, large scale deployment of smart meters is in progress. These devices enable detailed around-the-clock monitoring of energy usage. Specifically, each smart meter records accurately the electrical load for a given property at 10 second intervals, 24 hours a day. This granular data captures detailed habits and routines through user interactions with electrical devices. The research presented in this thesis exploits this infrastructure by using a novel approach that addresses the limitations associated with current Ambient Assistive Living technologies. By applying a novel load disaggregation technique and leveraging both machine learning and cloud computing infrastructure, a comprehensive, nonintrusive and personalised solution is achieved. This is accomplished by correlating the detection of individual electrical appliances and correlating them with an individual’s Activities of Daily Living. By utilising a random decision forest, the system is able to detect the use of 5 appliance types from an aggregated load environment with an accuracy of 96%. By presenting the results as vectors to a second classifier both normal and abnormal patient behaviour is detected with an accuracy of 92.64% and a mean squared error rate of 0.0736 using a random decision forest. The approach presented in this thesis is validated through a comprehensive patient trial, which demonstrates that the detection of both normal and abnormal patient behaviour is possible.

Primary headache disorders are the most common complaints worldwide. The socioeconomic and personal impact of headache disorders is enormous, as it is the leading cause of workplace absence. Headache patients’ consultations are increasing as the population has increased in size, live longer and many people have multiple conditions, however, access to specialist services across the UK is currently inequitable because the numbers of trained consultant neurologists in the UK are 10 times lower than other European countries. Additionally, more than two third of headache cases presented to primary care were labelled with unspecified headache. Therefore, an alternative pathway to diagnose and manage patients with primary headache could be crucial to reducing the need for specialist assessment and increase capacity within the current service model. Several recent studies have targeted this issue through the development of clinical decision support systems, which can help non-specialist doctors and general practitioners to diagnose patients with primary headache disorders in primary clinics. However, the majority of these studies were following a rule-based system style, in which the rules were summarised and expressed by a computer engineer. This style carries many downsides, and we will discuss them later on in this dissertation. In this study, we are adopting a completely different approach. The use of machine learning is recruited for the classification of primary headache disorders, for which a dataset of 832 records of patients with primary headaches was considered, originating from three medical centres located in Turkey. Three main types of primary headaches were derived from the data set including Tension Type Headache in both episodic and chronic forms, Migraine with and without Aura, followed by Trigeminal Autonomic Cephalalgia that further subdivided into Cluster headache, paroxysmal hemicrania and short-lasting unilateral neuralgiform headache attacks with conjunctival injection and tearing. Six popular machine-learning based classifiers, including linear and non-linear ensemble learning, in addition to one regression based procedure, have been evaluated for the classification of primary headaches within a supervised learning setting, achieving highest aggregate performance outcomes of AUC 0.923, sensitivity 0.897, and overall classification accuracy of 0.843. This study also introduces the proposed HydroApp system, which is an M-health based personalised application for the follow-up of patients with long-term conditions such as chronic headache and hydrocephalus. We managed to develop this system with the supervision of headache specialists at Ashford hospital, London, and neurology experts at Walton Centre and Alder Hey hospital Liverpool. We have successfully investigated the acceptance of using such an M-health based system via an online questionnaire, where 86% of paediatric patients and 60% of adult patients were interested in using HydroApp system to manage their conditions. Features and functions offered by HydroApp system such as recording headache score, recording of general health and well-being as well as alerting the treating team, have been perceived as very or extremely important aspects from patients’ point of view. The study concludes that the advances in intelligent systems and M-health applications represent a promising atmosphere through which to identify alternative solutions, which in turn increases the capacity in the current service model and improves diagnostic capability in the primary headache domain and beyond.

The growing demand for ever more greener aero engines has led to ever more challenging designs and higher quality products. An investigation into Coordinate Measuring Machine measurement uncertainty using physical measurements and virtual simulations revealed that there were several factors that can affect the measurement uncertainty of a specific task. Measurement uncertainty can be affected by temperature, form error and measurement strategy as well as Coordinate Measuring Machine specification. Furthermore the sensitivity of circular features size and position varied, when applying different substitute geometry algorithms was demonstrated. The Least Squares Circle algorithm was found to be more stable when compared with the Maximum Inscribed Circle and the Minimum Circumscribed Circle. In all experiments it was found that the standard deviation when applying Least Squares Circle was of smaller magnitude but similar trends when compared with Maximum Inscribed Circle and the Minimum Circumscribed Circle. A Virtual Coordinate Measuring Machine was evaluated by simulating physical measurement scenarios of different artefacts and different features. The results revealed good correlation between physical measurements uncertainty results and the virtual simulations. A novel methodology for the automated assessment of leading edge airfoil profiles was developed by extracting the curvature of airfoil leading edge, and the method lead to a patent where undesirable features such as flats or rapid changes in curvature could be identified and sentenced. A software package named Blade Inspect was developed in conjunction with Aachen (Fraunhoufer) University for the automated assessment and integrated with a shop floor execution system in a pre-production facility. The software used a curvature tolerancing method to sentence the leading edge profiles which aimed at removing the subjectivity associated with the manual vision inspection method. Initial trials in the pre-production facility showed that the software could sentence 200 profiles in 5 minutes successfully. This resulted in a significant improvement over the current manual visual inspection method which required 3 hours to assess the same number of leading edge profiles.

Although historical studies are frequently perceived as clear narratives defined by a series of fixed events; in reality, even where critical historical events may be identified, historic documentation frequently lacks corroborative detail to support verifiable interpretation. Consequently, interpretation rarely rises above the level of unproven assertion and is rarely tested against a range of evidence. Agent-based simulation can provide an opportunity to break these cycles of academic claim and counter-claim. This thesis discusses the development of an agent-based simulation designed to investigate medieval military logistics so that new evidence may be generated to supplement existing historical analysis. It uses as a case-study the Byzantine army’s march to the battle of Manzikert (AD 1071), a key event in medieval history. It describes the design and implementation of a series of agent-based models and presents the results of these models. The analysis of these results shows that agent-based modelling is a powerful tool in investigating the practical limitations faced by medieval armies on campaign.

Verification of access control systems against vulnerabilities has always been a challenging problem in the world of computer security. The complication of security policies in large- scale multi-agent systems increases the possible existence of vulnerabilities as a result of mistakes in policy definition. This thesis explores automated methods in order to verify temporal and epistemic properties of access control systems. While temporal property verification can reveal a considerable number of security holes, verification of epistemic properties in multi-agent systems enable us to infer about agents' knowledge in the system and hence, to detect unauthorized information flow. This thesis first presents a framework for knowledge-based verification of dynamic access control policies. This framework models a coalition-based system, which evaluates if a property or a goal can be achieved by a coalition of agents restricted by a set of permissions defined in the policy. Knowledge is restricted to the information that agents can acquire by reading system information in order to increase time and memory efficiency. The framework has its own model-checking method and is implemented in Java and released as an open source tool named \(\char{cmmi10}{0x50}\)\(\char{cmmi10}{0x6f}\)\(\char{cmmi10}{0x6c}\)\(\char{cmmi10}{0x69}\)\(\char{cmmi10}{0x56}\)\(\char{cmmi10}{0x65}\)\(\char{cmmi10}{0x72}\). In order to detect information leakage as a result of reasoning, the second part of this thesis presents a complimentary technique that evaluates access control policies over temporal-epistemic properties where the knowledge is gained by reasoning. We will demonstrate several case studies for a subset of properties that deal with reasoning about knowledge. To increase the efficiency, we develop an automated abstraction refinement technique for evaluating temporal-epistemic properties. For the last part of the thesis, we develop a sound and complete algorithm in order to identify information leakage in Datalog-based trust management systems.

In order to study computationally increasingly complex systems using theoretical methods model Hamiltonians are required to accurately describe the potential energy surface they represent. Also ab-initio methods improve the calculation of the excited states of these complex systems becomes increasingly feasible. One such model Hamiltonian described herein, the Vibronic Coupling Hamiltonian, has previously shown its versitility and ability to describe a variety of non-adiabatic problems. This thesis describes a new method, a genetic algorithm, for the parameterisation of the Vibronic Coupling Hamiltonian to describe both previously calculated potential energy surfaces (allene and pentatetraene) and newly calculated (cyclo-butadiene and toluene) potential energy surfaces. In order to test this genetic algorithm quantum nuclear dynamics calculations were performed using the multi-configurational time dependent hartree method and the results compared to experiment.