Browsing by Title

The fast development cycles of Web User Interfaces (UI) create challenges for test automation to keep up with the changes in web elements. Test automation may suffer from test breakages after developers update the UIs of the System Under Test (SUT). Test breakages are not defects or bugs in the SUT, but a failure in test automation code. Failing to correctly locate web element from the UI, is one of the key reasons for test breakages to occur. Prior work to gain self-healing of element locators, has been traditionally done with different algorithms and recently with the help of Large Language Models (LLMs). This thesis aims to discover how to enable self-healing locators for Robot Framework Web UI tests, are there some web element locator types that are more easily repaired than others, and which LLMs should be used for this task. An experimental study was conducted for enabling self-healing locators for Robot Framework. Custom Robot Framework library was created with Python, which was tested for eight different locator strategies raised from locator breakage taxonomy. Results show that the best performance in self-healing locators is gained by using the bigger LLMs. GPT-4 Turbo and Mistral Large showed the best performance accuracy by repairing 87,5% of the locators in the Robot Framework test cases. The worst performer was Mistral 7B Instruct which was not able to correct any locators. Using LLMs for self-healing locators in Robot Framework tests is possible. To get the best results for self-healing locators, my results suggest that practitioners should focus on LLM prompt designing, in the usage of candidate algorithm with locator version history and use the biggest LLMs available if possible.

Many diseases are dependant of the concentration of hormones in our body, especially for women. The risk of having this type of disease such as sex-related cancer is lower in Asian countries thanks to the high consumption of soy in these countries. Indeed, soy contains several isoflavones such as daidzein or genistein which act as phytoestrogens once metabolized in our body. It is due to the fact that they are similar to estrogens but they do not have or have a lower biological effect than estrogens compounds. (S)-dihydrodaidzein, a product of the metabolism of daidzein which is one of these isoflavones, has a stronger bioactivity than its parent compound but only 30% of the population in the world can obtain it naturally. The current research was focused on the stereoselective synthesis of dihydrodaidzein by hydrogenation using a chiral catalyst that will favor the obtention of only one enantiomer. (S)-proline was used as the chiral modifier. Two types of hydrogenation were used, transfer hydrogenation and hydrogenation under pressure, the latter being the one where we obtained the maximum of enantiomeric excess. 44% was achieved after 3 hours of hydrogenation with palladium on activated carbon as catalyst at 20 bars of pressure of dihydrogen. It is the first report of the use of (S)-proline on another compound than isophorone and we obtained higher enantiomeric excess for the same conditions. This study relies on the use of one and two-dimensional NMR, preparative HPLC-MS and chiral HPLC to identify and to obtain pure compounds and the enantiomeric excess of dihydrodaizein. Future research should focus on the kinetic study of this reaction and the elucidation of the mechanism.

Propargyl or methacrylate end-functionalized polylactides are important intermediates in polymer synthesis towards their application in the biomedical field. Through these intermediates hydrophobic polylactides are post polymerized with hydrophilic monomers or coupled with preformed hydrophilic homopolymers to obtain amphiphilic copolymers that possess qualities that they wouldn’t otherwise have. These systems show a unique property: self-assembly into micellar structures that can be utilized in drug delivery applications. Polylactides (PLA) offer biocompatibility and biodegradability to its non-toxic and non-carcinogenic metabolites for these biomedical applications. Understanding the synthesis, modification and processing of PLA in this light is a cornerstone for successful development of new PLA-based biomaterials. This thesis gives an overview of polylactide end functionalization, their post polymerization and documents attempts to synthesize different end functional polylactides. The experimental part of this thesis focuses on the synthesis of propargyl and methacrylate end-functionalized PLA:s with different chain lengths and their characterization with the appropriate polymer characterization techniques.

Kasvavan kaupungistumisen johdosta kaupunki-ilmaston tutkiminen on kasvanut selvästi viimeisten vuosikymmenien aikana. Yleisimmin tunnettu kaupungistumisen aiheuttama ilmiö on lämpösaarekeilmiö, jossa yöaikaan kaupungin ilmanlämpötila pysyttelee selvästi ympäröiviä maaseutuja korkeammalla. Tämä ja monet muut kaupungissa koetut ilmiöt ovat ihmisen aiheuttamia ja siksi on tärkeää selvittää, miten kasvava kaupungistuminen tulee jatkossa vaikuttamaan meidän jokapäiväiseen elämään, terveyteen ja ympäristöömme. Helsingin Kumpulassa sijaitseva SMEAR III (Station for Measuring Ecosystem-Atmosphere Relationships) kaupunkimittausasema sai syyskuussa 2010 uuden mittauspisteen Helsingin keskustaan Hotelli Tornin huipulle, missä mitataan mm. turbulenttista latentin ja havaittavan lämmön vuota. Tässä tutkielmassa tarkastellaan kaupunkialueen energiatasapainoa Helsingin keskustassa Tornissa tehtävien mittauksien avulla, sekä perehdytään energiatasapainon mallintamiseen yhdellä maailman pohjoisimmista kaupunkimittausasemista. Tutkielmassa käytetään mittauksia, jotka kattavat kokonaisen vuoden marraskuusta 2010 lokakuuhun 2011. Mitattuja lämmön turbulenttisia voita verrataan Surface Urban Energy and Water Balance Scheme -mallilla (SUEWS) mallinnettuihin voihin. Tarkoitus on selvittää sitä, kuinka hyvin malli toimii Helsingin keskustan hyvin tiheään rakennetulla kaupunkialueella. SUEWS:n lisäksi työssä käytetään Kormannin ja Meixnerin lähdealuemallia, jolla saadaan lasketuksi arvio mitattujen voiden lähdealueista. Saatujen tuloksien mukaan Helsingin keskustassa havaittavan lämmön vuo on monta kertaa latentin lämmön vuota suurempaa. Mallinnettaessa energiatasapainon komponentteja SUEWS-mallilla havaittavan lämmön vuon arvot olivat melko lähellä mitattuja arvoja, mutta latentin lämmön vuon tapauksessa mallinnetut arvot olivat liian pieniä. Ongelmia aiheutti saatavilla olevan energian ja kosteuden liian vähäinen määrä, jotka johtuivat mm. mallinnuksessa käytetystä liian suuresta albedosta ja maankäytön aineiston karkeasta resoluutiosta. Antropogeenisen lämmön vuon määrä Helsingissä oli mallinnuksen mukaan noin 20-50 W m-2:llä, joka vastaa lauhkean vyöhykkeen kaupunkien keskiarvoja. Varastotermin käyttäytyminen oli oikeanlaista, jossa päivällä energiaa sitoutuu rakenteisiin ja yöllä sitä vapautuu ilmaan turbulenttisina voina. Yhteenlasketut energian lähdetermit eivät kuitenkaan olleet tarpeeksi isoja, joka kävi hyvin selväksi energiatasapainosuhteen avulla tehdystä vertailusta, jossa suhteeksi saatiin 1,26.

Energy efficiency is one of the key factors impacting the green behavior and operational expenses of telecommunication core network operations. This thesis study is aimed for finding out possible technique to reduce energy consumption in telecommunication infrastructure nodes. The study concentrates on traffic management operation (e.g. media stream control, ATM adaptation) within network processors [LeJ03], categorized as control plane. The control plane of the telecommunication infrastructure node is a custom built high performance cluster which consists of multiple GPPs (General Purpose Processor) interconnected by high-speed and low-latency network. Due to application configurations in particular GPP unit and redundancy issues, energy usage is not optimal. In this thesis, our approach is to gain elastic capacity within the control plane cluster to reduce power consumption. This scales down and wakes up certain GPP units depending on traffic load situations. For elasticity, our study moves toward the virtual machine (VM) migration technique in the control plane cluster through system virtualization. The traffic load situation triggers VM migration on demand. Virtual machine live migration brings the benefit of enhanced performance and resiliency of the control plane cluster. We compare the state-of-the-art power aware computing resource scheduling in cluster-based nodes with VM migration technique. Our research does not propose any change in data plane architecture as we are mainly concentrating on the control plane. This study shows, VM migration can be an efficient approach to significantly reduce energy consumption in control plane of cluster-based telecommunication infrastructure nodes without interrupting performance/throughput, while guaranteeing full connectivity and maximum link utilization.

The performance of today's smartphones is heavily limited by their batteries. To meet the increasing demands from end-users on smartphones with better battery life, Android mobile manufacturers are hunting for new approaches to improve their phones' batteries. For example, they use smart battery interfaces on their existing batteries, and they also render more usable power-saving features and settings that mobile users can use to minimize the power consumption of smartphones. But unfortunately, users always have inadequate knowledge about power characteristics and features of their mobile devices. This highlights the necessity to let users know how they should deal with limited battery lifetime and thus perform the optimal and effective energy management. Especially when the battery gets critically low, users should have a good understanding of how to better manage the remaining battery and take appropriate measures to make their phones last longer for at least keeping basic email and call communication. The objective of this thesis is to provide Android mobile users with a series of control points on maximizing the lifetime of their low-battery smartphones when they desire to maintain basic communication for emailing and calling. To achieve this goal, a detailed analysis will be performed on an Android smartphone, Samsung Google Nexus S. In the beginning, I will introduce an overview of energy profiling on the Android platform, where Carat will be presented as a practical example of energy profilers aiming for detecting energy bugs on users' smartphones and providing them with diagnosis reports to reduce the power consumed by their phones. After that, I attempt to find out how general hardware subsystems (eg. screen backlight and internet connections) and background application processes will impact battery life of smartphones by undertaking physical power measurements of the tested phone under a wide range of realistic usage scenarios. At last, I will evaluate the experimental results from users' perspectives to figure out under which circumstances battery life of smartphones will be maximized, and then based on the obtained findings I will design actionable recommendations for users on battery management. From this study, I have found that screen backlight, Internet connections and background application processes indeed impact battery life when users are emailing and calling through low-battery smartphones. In addition, to improve the lifespan of smartphones used for emailing, users should connect to WiFi, adjust screen backlight to the minimum level and stop running those application processes in the background, which have been identified by Carat as energy hogs or bugs. Moreover, if users want to have Internet calls through the Skype, they are recommended to choose WiFi connections and kill background application processes. At the same time, if Dialer is used for calling, users should disable any Internet connections to reduce the power consumption of their smartphones.

Computational aspects of argumentation are a central research topic of modern artificial intelligence. A core formal model for argumentation, where the inner structure of arguments is abstracted away, was provided by Dung in the form of abstract argumentation frameworks (AFs). AFs are syntactically directed graphs with the nodes representing arguments and edges representing attacks between them. Given the AF, sets of jointly acceptable arguments or extensions are defined via different semantics. The computational complexity and algorithmic solutions to so-called static problems, such as the enumeration of extensions, is a well-studied topic. Since argumentation is a dynamic process, understanding the dynamic aspects of AFs is also important. However, computational aspects of dynamic problems have not been studied thoroughly. This work concentrates on different forms of enforcement, which is a core dynamic problem in the area of abstract argumentation. In this case, given an AF, one wants to modify it by adding and removing attacks in a way that a given set of arguments becomes an extension (extension enforcement) or that given arguments are credulously or skeptically accepted (status enforcement). In this thesis, the enforcement problem is viewed as a constrained optimization task where the change to the attack structure is minimized. The computational complexity of the extension and status enforcement problems is analyzed, showing that they are in the general case NP-hard optimization problems. Motivated by this, algorithms are presented based on the Boolean optimization paradigm of maximum satisfiability (MaxSAT) for the NP-complete variants, and counterexample-guided abstraction refinement (CEGAR) procedures, where an interplay between MaxSAT and Boolean satisfiability (SAT) solvers is utilized, for problems beyond NP. The algorithms are implemented in the open source software system Pakota, which is empirically evaluated on randomly generated enforcement instances.

Geothermal energy is a growing industry and with Enhanced Geothermal System (EGS) technology it is possible to utilize geothermal energy in low heat flow areas. The ongoing EGS project in Southern Finland provides a great opportunity to learn and explore EGS technologies in a complex environment: hard crystalline rock, high pressure and low hydraulic permeability. This work describes physics behind an EGS plant, as well as basic concept of EGS, give examples of some existing plants and make calculations of how much power a plant in Finland can produce. In order to plan and build a successful plant, suitable parameters for the system are determined by modelling. The modelling is done analytically and numerically. Physical properties governing the EGS models are conductive and convective heat transfer and rock hydraulic properties that allow fluid flow. Hydraulic permeability is discussed in detail, because it is the key parameter in EGS: rock is stimulated in order to enhance permeability in order to make fluid flow possible through interconnected fractures. It is a spatially correlated parameter and it is distributed lognormally making fluid flow highly channelled. Modelling of heat and mass transfer aims to parametrize an EGS plant in the conditions of Southern Finland. The parameters governing heat transfer with fluid flowing in the geothermal reservoir are size of the reservoir and fluid velocity, which depends on matrix permeability. The larger the reservoir the more hot contact area fluid encounters and the better it heats up, the slower the flow, the longer time fluid stays in the reservoir and therefore heats up more. High flow rates cool the reservoir rapidly. However, a large reservoir is difficult to achieve, maintaining enhanced permeability requires relatively high fluid flow rates and the higher the flow rate, the more power the plant produces, so slow flow is not economically feasible. Analytical models are done with Matlab and numerical models are done with finite-element software COMSOL Multiphysics. Numerical models benchmark the analytical solutions and use spatially correlated permeability to modify fluid flow pattern and see how temperature in the reservoir changes with changes in fluid flow. The results show that creating large reservoir that could operate for 20 years with desired power production is unrealistic. Total output fluid flow required to produce over 1 MW of power is 10 kg/s. At such rate there is a risk that the reservoir cools and output fluid temperature is not sufficient for power production. In case of heterogeneous permeability connectivity of the reservoir is not as good as in case of homogeneous permeability and there is a risk that total fluid flow in the reservoir is slower and therefore less power produced.

Accurate interpretation of high-resolution molecule spectral data is important for scientific research such as atmospheric chemistry. This master thesis describes the development of a user-friendly visualization tool to improve the accessibility and interpretation of spectral data in the HITRAN database. By using Tkinker, a Python interface for creating graphical user interfaces (GUI), the spectral simulation tool simplifies the data visualization and analysis of molecular spectra. Users can plot line intensities and absorption spectra of various molecular species and isotopes, and adjust parameters such as wavelength, wavenumber, frequency, temperature, pressure, length, and volume mixing ratio (VMR). The GUI allows for the selection of linear or logarithmic scales to improve the clarity and depth of the spectral analysis. The GUI not only provides a practical application for the visualization of complex spectral data, but also contributes to expanding the accessibility of the HITRAN database, making it more accessible to researchers, professionals, and students in related fields. In conclusion, the thesis describes the related background and theory of the tool, the technical implementation of the GUI and its validation, a case study on NH3 measurements, and the potential for future improvements.

Denna avhandling ger en introduktion till fraktal geometri utgående från klassiska exempel. Avsikten är att ge en mångsidig inblick i vad fraktaler är och vilka nya geometriska begrepp som behövs för att beskriva dem. I brist på en rigorös definition beskrivs fraktaler utgående från deras typiska egenskaper, såsom exakt eller ungefärlig självlikhet och detaljerad struktur på varje nivå. Enklare uttryckt kommer ingen godtyckligt liten del av en fraktal figur att likna en linje och varje godtyckligt liten del av fraktalen innehåller delar som är exakta eller ungefärliga miniatyrer av hela mängden. Begreppet fraktal myntades av Benoit Mandelbrot år 1975 och blev populärt i samband med möjligheten att rita bilder av dem med datorer. Mycket av den matematik som används inom fraktal geometri utvecklades däremot under början av 1900-talet av Felix Hausdorff och Gaston Julia med flera. Två av de grundläggande begreppen i fraktal geometri är Hausdorff- och Minkowskidimension, vilka är generaliseringar av begreppet dimension. För fraktaler är dessa i allmänhet inte heltal. Den här introduktionen till fraktal geometri omskriver deras definitioner och egenskaper samt olika metoder för att beräkna dem för ett antal variarande exempel. Avhandlingen behandlar också den rigorösa konstruktionen av itererande funktionssystem (IFS) vars attraktorer oftast är fraktaler. IFS ger en metod för att systematiskt konstruera och undersöka en stor mängd fraktaler. Avhandlingen beskriver också hur fraktaler uppstår i mycket olika kontexter såsom talteori, fraktal interpolation av data och komplex dynamik. Gemensamt för alla fraktaler är den roll som oändlig rekursion eller oändliga iterationer spelar i fraktalernas konstruktion. Fraktalers koppling till kaos behandlas också ytligt. Avslutningsvis diskuteras vilken koppling fraktal geometri har till naturen och vilken nytta vi eventuellt kan ha av området. Avhandlingen är ämnad för läsare med olika nivå av matematiska förkunskaper och innehåller både lättare och svårare koncept.