Schack är ett spel som ofta används för att undersöka olika tekniker inom artificiell intelligens (AI). I det här arbetet ställs frågan om det går att utveckla en AI-agent vars beslutsfattande är baserat på tekniken Case-based Reasoning (CBR) med grundlig likhet som spelar bättre med fallbaser baserade på bättre experter. En AI-agent har utvecklats som spelat ett antal partier mot sig själv med olika fallbaser baserade på olika experter. Efter att ha undersökt resultatet visade de sig att AI-agenten spelar så dåligt att den nästan aldrig lyckades vinna oavsett fallbas, vilket gjorde att det inte gick att rangordna dem efter skicklighet. I framtida arbete är det intressant att undersöka andra likheter än grundlig likhet. Det är även av intresse att undersöka om en CBR-baserad schackspelande AI-agent kan spela schack med hög skicklighet.

The project introduces and subsequently proposes a novel framework, based on the medical model of diagnosis, to facilitate business students' preparation of case analysis. The project also shows students how to apply the framework in a sample case study.

This thesis is about symptom-based troubleshooting of heavy vehicles. The existing troubleshooting system at Scania is adapted to handle errors based on electronic fault codes. This means that some faults, such as mechanical faults when sensors are missing, are difficult to troubleshoot. In the thesis, a method is developed that will be a part of a symptom-based troubleshooting system which can handle all types of errors. The main objectives of the thesis are both to develop a method that can link customer symptoms with system reactions and also to develop formats for both customer symptoms and FMEA for the developed method. In the thesis, a literature study was first conducted in which troubleshooting methods and principles for the formalization of customer data were identified. The identified troubleshooting methods were Bayesian Network, Case-Based-Reasoning and Fault tree analysis. A case study was then conducted which was based on several documents for troubleshooting in gas engines and gas tanks. In the case study, data from the literature study and the empirically collected data were used to develop the final concept of the method. The case study included, among other things, semi-structured interviews to map out the existing troubleshooting process, and a workshop to choose the final concept. In order to meet the objectives of the thesis two research questions and one question linked to the case study were formulated: Research Questions: • RQ1: How is the troubleshooting process affected by the methods that can be used to link customer symptoms with system reactions in heavy vehicles? • RQ2: How can customer data and FMEA be formalized in order to be useful in the troubleshooting process of heavy vehicles? Case Study: • What kind of data is missing from Scania’s existing documentation to link customer symptoms with system reactions? The thesis resulted in a method based on two troubleshooting methods Bayesian network and Case-Based-Reasoning. The method links customer symptoms with system reactions by excluding human considerations and instead relying on previously documented cases and probabilities. A requirement for using this method is a cooperation between customer support, mechanics and development engineers. The formalization of customer symptoms in the developed method is based on what good data is for mechanics in troubleshooting contexts and what customers are capable of communicating; deviation – the customer’s description of the vehicle’s unexpected condition, position – where the customer considers the deviation to be present, context – what happened before, during and after the deviation was discovered. The conclusions that can be drawn is that it is not necessary to link customer symptoms with system reactions since the developed method allows the customer symptoms to be linked directly to the corrective actions needed. In addition, it was noted that the existing documentation at Scania on customer symptoms and system reactions is insufficient. However, this is not problematic as it was shown that FMEA is redundant for the method developed. In order for customer data to be useful, the formalization should include deviation, position and context. Further conclusions are that the role of the customer support becomes less critical when data driven troubleshooting methods are used, and that the accuracy of the developed method will improve over time as more data will be collected. / Detta arbete behandlar symptombaserad felsökning av tunga fordon. Scanias befintliga felsökningssystem är anpassat för att hantera fel som grundas i elektroniska felkoder. Detta innebär att vissa typer av fel, såsom mekaniska fel när sensorer saknas, är svåra att felsöka. I detta arbete utvecklas en metod som ska ingå i ett symptombaserat felsökningssystem eftersom ett sådant system kan hantera alla typer av fel. Målen med arbetet är att utveckla en metod som kan sammankoppla kundsymptom med systemreaktioner, och utveckla format för kundsymptom och FMEA för den framtagna metoden. I arbetet utfördes först en litteraturstudie där felsökningsmetoder och principer för formaliseringen av kunddata identifierades. Felsökningsmetoderna som identifierades var Bayesiska nätvkerk, Case-Based-Reasoning och Felträdsanalys. Därefter utfördes en fallstudie som grundades på underlag om felsökning inom gasmotorer och gastankar. I fallstudien användes data från litteraturstudien och den empiriskt insamlade data för att utveckla det slutgiltiga konceptet. I fallstudien utfördes bland annat semistrukturerade intervjuer för att kartlägga den befintliga felsökningsprocessen, och en workshop för att kunna välja det slutgiltiga konceptet. För att kunna uppfylla arbetets mål formulerades två forskningsfrågor och en frågeställning kopplad till fallstudien: Forskningsfrågor: • F1: Hur påverkas felsökningsprocessen utifrån de metoder som kan användas för att sammankoppla kundsymptom med systemreaktioner inom tunga fordon? • F2: Hur kan kunddata och FMEA formaliseras för att vara användbara inom felsökningsprocessen av tunga fordon? Fallstudie: • Vilken data saknas i Scanias befintliga dokumentation för att kunna sammankoppla kundsymptom med systemreaktioner? Arbetet resulterade i en metod som baseras på de två felsökningsmetoderna Bayesiska nätverk och Case-Based-Reasoning. Metoden sammankopplar kundsymptom med systemreaktioner genom att exkludera mänskligt avvägande och istället förlita sig på tidigare dokumenterade fall och sannolikhet. En förutsättning för att metoden ska kunna användas är ett samarbete mellan kundmottagare, mekaniker och utvecklingsingenjörer. Formaliseringen av kundsymptom i den framtagna metoden bygger på vad bra data är för mekaniker i felsökningssammanhang och vad kunderna är kapabla att förmedla; avvikelse – kundens beskrivning av fordonets oväntade tillstånd, position – var anser kunden att avvikelsen förekommer, kontext – vad hände innan, under och efter att avvikelsen upptäcktes. Slutsatserna som kan dras utifrån arbetet är att det inte är nödvändigt att sammankoppla kundsymptom med systemreaktioner, utan kundsymptom kan sammankopplas direkt med åtgärder med den framtagna metoden. Dessutom noterades det att den befintliga dokumentationen hos Scania angående kundsymptom och systemreaktioner är bristfällig. Detta är inte problematiskt då det påvisades att FMEA inte är nödvändig för att metoden ska fungera. För att kunddata ska vara användbart bör formaliseringen ske med avvikelse, position och kontext. Ytterligare slutsatser är att kundmottagarrollen blir mindre kritisk när datadrivna felsökningsmetoder används, och att den framtagna metodens träffsäkerhet kommer att förbättras över tid allt eftersom mer data har samlats in.

This dissertation provides the theory, design, and implementation of a cognitive engine, the enabling technology of cognitive radio. A cognitive radio is a wireless communications device capable of sensing the environment and making decisions on how to use the available radio resources to enable communications with a certain quality of service. The cognitive engine, the intelligent system behind the cognitive radio, combines sensing, learning, and optimization algorithms to control and adapt the radio system from the physical layer and up the communication stack. The cognitive engine presented here provides a general framework to build and test cognitive engine algorithms and components such as sensing technology, optimization routines, and learning algorithms. The cognitive engine platform allows easy development of new components and algorithms to enhance the cognitive radio capabilities. It is shown in this dissertation that the platform can easily be used on a simulation system and then moved to a real radio system. The dissertation includes discussions of both theory and implementation of the cognitive engine. The need for and implementation of all of the cognitive components is strongly featured as well as the specific issues related to the development of algorithms for cognitive radio behavior. The discussion of the theory focuses largely on developing the optimization space to intelligently and successfully design waveforms for particular quality of service needs under given environmental conditions. The analysis develops the problem into a multi-objective optimization process to optimize and trade-of of services between objectives that measure performance, such as bit error rate, data rate, and power consumption. The discussion of the multi-objective optimization provides the foundation for the analysis of radio systems in this respect, and through this, methods and considerations for future developments. The theoretical work also investigates the use of learning to enhance the cognitive engine's capabilities through feed-back, learning, and knowledge representation. The results of this work include the analysis of cognitive radio design and implementation and the functional cognitive engine that is shown to work in both simulation and on-line experiments. Throughout, examples and explanations of building and interfacing cognitive components to the cognitive engine enable the use and extension of the cognitive engine for future work. / Ph. D.

published_or_final_version / Urban Planning and Environmental Management / Doctoral / Doctor of Philosophy

Diagrams are increasingly used in many design methods, and are being taught in a variety of contexts in higher education such as database conceptual design or software design in computer science. They are an important part of many assessments. Currently computer aided assessments are widely used for multiple choice questions. They lack the ability to assess a student's knowledge in a more comprehensive way, which is required for diagram-type student work. The aim of this research is to develop a semi-automatic assessment framework, which enables the use of computer to support the assessment process of diagrammatic solutions, with the focus of ensuring the consistency of grades and feedback on solutions. A novel trace model, that captures design traces of student solutions, was developed as a part of the framework and was used to provide the matching criteria for grouping the solutions. A new marking style, partial marking, was developed to mark these solution groups manually. The Case-Based Reasoning method is utilised in the framework to mark some of the groups automatically. A guideline for scenario writing was proposed to increase the efficiency of automatic marking. A prototype diagram editor, a marking tool and scenario writing environment were implemented for the proposed framework in order to demonstrate proof of concept. The results of experiments show that the framework is feasible to use in the formative assessment and it provides consistent marking and personalised feedback to the students. The framework also has the potential to significantly reduce the time and effort required by the examiner to mark student diagrams. Although the constructed framework was specifically used for the assessment of database diagrams, the framework is generic enough to be used for other types of graph-based diagram.

Planning for teaching imposes a significant burden on teachers, as teachers need to prepare different lesson plans for different classes according to various constraints. Statistical evidence shows that lesson planning in the Malaysian context is done in isolation and lesson plan sharing is limited. The purpose of this thesis is to investigate whether a case-based system can reduce the time teachers spend on constructing lesson plans. A case-based system was designed SmartLP. In this system, a case consists of a problem description and solution pair and an attributevalue representation for the case is used. SmartLP is a synthesis type of CBR system which attempts to create a new solution by combining parts of previous solutions in the adaptation. Five activities in the CBR cycle retrieve, reuse, revise, review and retain are created via three types of design: application, architectural and user interface. The inputs are the requirements and constraints of the curriculum and the student facilities available, and the output is the solution, i.e. appropriate elements of a lesson plan. The retrieval module consists of five types of search advanced search, hierarchical, Boolean, basic and browsing. Solving a problem in this system involves obtaining a problem description, measuring the similarity of the current problem to previous problems stored in a database, retrieving one or more similar cases and attempting to reuse the solution of the retrieved cases, possibly after adaptation. Case adaptation for multiple lesson plans helps teachers to customise the retrieved plan to suit their constraints. This is followed by case revision, which allows users to access and revise their constructed lesson plans in the system. Validation mechanisms, through case verification, ensure that the retained cases are of quality. A formative study was conducted to investigate the effects of SmartLP on performance. The study revealed that all the lesson plans constructed with SmartLP assistance took significantly less time than the control lesson plans constructed without SmartLP assistance, although they might have access to computers and other tools. No significant difference in writing quality, measured by a scoring system, was noticed for the control group, who constructed lesson plans on the same tasks without receiving any assistance. The limitations of SmartLP are indicated and the focus of further research is proposed. Keywords: Case-based system, CBR approach, knowledge acquisition, knowledge representation, case representation, evaluation, lesson planning.

Cette thèse propose deux approches pour améliorer la qualité des réponses d'un système de raisonnement à partir de cas (RàPC) utilisant des connaissances produites par une communauté en ligne. La première approche concerne la mise en œuvre d'un modèle permettant de gérer la fiabilité des connaissances produites par la communauté sous la forme d'un score. Ce score de fiabilité est utilisé d'une part pour filtrer les connaissances non fiables afin qu'elles ne soient pas utilisées par le système de RàPC et d'autre part pour classer les réponses retournées par le système. La deuxième approche concerne la représentation de la typicalité entre sous-classes et classes dans une organisation hiérarchique. La typicalité est alors utilisée pour réorganiser les connaissances hiérarchiques utilisées par le système de RàPC. L'apport de ces deux approches a été évalué dans le cadre de eTaaable, un système de RàPC qui adapte des recettes de cuisine en utilisant des connaissances produites par une communauté en ligne. L'évaluation montre que la gestion de la fiabilité des connaissances produites par la communauté améliore la qualité des réponses retournées par eTaaable. De même, l'évaluation montre que l'utilisation par eTaaable des hiérarchies des connaissances réorganisées en exploitant la typicalité améliore également la qualité des réponses / This research work presents two approaches to improve the quality of the results returned by a case-based reasoning system (CBR) exploiting knowledge produced by an e-community. The first approach relies on a new model to manage the trustworthiness of the knowledge produced by the e community. In this model, the trustworthiness is represented through a score which is used to filter untrustworthy knowledge so that the CBR system will not use it anymore. Moreover, the trustworthiness score is also used to rank the CBR results. The second approach addresses the issue of representing the typicality between subclasses and classes in a hierarchy. The typicality is used to change the hierarchical organization used by the CBR system. Both approaches have been evaluated in the framework of eTaaable, a CBR system which adapts cooking recipes using knowledge coming from an e-community. The evaluations show that managing the trustworthiness of the knowledge produced by an e-community improves the quality of the results returned by eTaaable. The evaluations also shows that eTaaable returns also better results when using knowledge reorganized according to typicality.

<p>Increased stress is a continuing problem in our present world. Especiallynegative stress could cause serious health problems if it remainsundiagnosed/misdiagnosed and untreated. In the stress medicine, clinicians’measure blood pressure, ECG, finger temperature and breathing rate during anumber of exercises to diagnose stress-related disorders. One of the physiologicalparameters for quantifying stress levels is the finger temperature that helps theclinicians in diagnosis and treatment of stress. However, in practice, it is difficultand tedious for a clinician to understand, interpret and analyze complex, lengthysequential sensor signals. There are only few experts who are able to diagnose andpredict stress-related problems. A system that can help the clinician in diagnosingstress is important, but the large individual variations make it difficult to build sucha system.This research work has attempted to investigate several artificial Intelligencetechniques to develop an intelligent, integrated sensor system for diagnosis andtreatment plan in the Psychophysiological domain. To diagnose individualsensitivity to stress, case-based reasoning is applied as a core technique to facilitateexperience reuse by retrieving previous similar cases. Further, fuzzy techniques arealso employed and incorporated into the case-based reasoning system to handlevagueness, uncertainty inherently existing in clinicians reasoning process. Thevalidation of the approach is based on close collaboration with experts andmeasurements from twenty four persons used as reference.Thirty nine time series from these 24 persons have been used to evaluate theapproach (in terms of the matching algorithms) and an expert has ranked andestimated similarity which shows a level of performance close to an expert. Theproposed system could be used as an expert for a less experienced clinician or as asecond option for an experienced clinician to their decision making process.</p> / Integrated Personal Health Optimizing System (IPOS)

<p>The history of oil well engineering applications has revealed that the frequent operational problems are still common in oil well practice. Well blowouts, stuck pipes, well leakages are examples of the repeated problems in the oil well engineering industry. The main reason why these unwanted problems are unavoidable can be the complexity and uncertainties of the oil well processes. Unforeseen problems happen again and again, because they are not fully predictable, which could be due to lack of sufficient data or improper modelling to simulate the real conditions in the process. Traditional mathematical models have not been able to totally eliminate unwanted oil well problems because of the many involved simplifications, uncertainties, and incomplete information. This research work proposes a new approach and breakthrough for overcoming these challenges. The main objective of this study is merging two scientific fields; artificial intelligence and petroleum engineering in order to implement a new methodology.</p><p>Case-Based Reasoning (CBR) and Model-Based Reasoning (MBR), two branches of the artificial intelligence science, are applied for solving complex oil well problems. There are many CBR and MBR modelling tools which are generally used for different applications for implementing and demonstrating CBR and MBR methodologies; however, in this study, the Creek system which combines CBR and MBR has been utilized as a framework. One specific challenging task related to oil well engineering has been selected to exemplify and examine the methodology. To select a correct candidate for this application was a challenging step by itself. After testing many different issues in the oil well engineering, a well integrity issue has been chosen for the context. Thus, 18 leaking wells, production and injection wells, from three different oil fields have been analysed in depth. Then, they have been encoded and stored as cases in an ontology model given the name Wellogy.</p><p>The challenges related to well integrity issues are a growing concern. Many oil wells have been reported with annulus gas leaks (called internal leaks) on the Norwegian Continental Shelf (NCS) area. Interventions to repair the leaking wells or closing and abandoning wells have led to: high operating cost, low overall oil recovery, and in some cases unsafe operation. The reasons why leakages occur can be different, and finding the causes is a very complex task. For gas lift and gas injection wells the integrity of the well is often compromised. As the pressure of the hydrocarbon reserves decreases, particularly in mature fields, the need for boosting increases. Gas is injected into the well either to lift the oil in the production well or to maintain pressure in the reservoir from the injection well. The challenge is that this gas can lead to breakdown of the well integrity and cause leakages. However, as there are many types of leakages that can occur and due to their complexity it can be hard to find the cause or causal relationships. For this purpose, a new methodology, the Creek tool, which combines CBR and MBR is applied to investigate the reasons for the leakages. Creek is basically a CBR system, but it also includes MBR methods.</p><p>In addition to the well integrity cases, two complex cases (knowledge-rich cases) within oil well engineering have also been studied and analysed through the research work which is part of the PhD. The goal here is to show how the knowledge stored in two cases can be extracted for the CBR application.</p><p>A model comprising general knowledge (well-known rules and theories) and specific knowledge (stored in cases) has been developed. The results of the Wellogy model show that the CBR methodology can automate reasoning in addition to human reasoning through solving complex and repeated oil well problems. Moreover, the methodology showed that the valuable knowledge gained through the solved cases can be sustained and whenever it is needed, it can be retrieved and reused. The model has been verified for unsolved cases by evaluating case-matching results. The model gives elaborated explanations of the unsolved cases through the building of causal relationships. The model also facilitates knowledge acquisition and learning curves through its growing case base.</p><p>The study showed that building a CBR model is a rather time-consuming process due to four reasons:</p><p>1. Finding appropriate cases for the CBR application is not straightforward</p><p>2. Challenges related to constructing cases when transforming reported information to symbolic entities</p><p>3. Lack of defined criteria for amount of information (number of findings) for cases</p><p>4. Incomplete data and information to fully describe problems of the cases at the knowledge level</p><p>In this study only 12 solved cases (knowledge-rich cases) have been built in the Wellogy model. More cases (typically hundreds for knowledge-lean cases and around 50 for knowledge-rich cases) would be required to have a robust and efficient CBR model. As the CBR methodology is a new approach for solving complex oil well problems (research and development phase), additional research work is necessary for both areas, i.e. developing CBR frameworks (user interfaces) and building CBR models (core of CBR). Feasibility studies should be performed for implemented CBR models in order to use them in real oil field operations. So far, the existing Wellogy model has showed some benefits in terms of; representing the knowledge of leaking well cases in the form of an ontology, retrieving solved cases, and reusing pervious cases.</p>