Transportation is embedded in the fabric of society and a key enabler of socio-economic development, but it is also a major source of carbon dioxide (CO2) and local air pollution (AP). Cities collectively account for around three quarters of total energy-related CO2 emissions, and the negative health impacts from local APs are most felt in dense urban environments. Thus, transitioning away from current fossil fuel regime in urban transportation is necessary to address both global and local challenges. Mathematical models as energy system optimisation models (ESOMs) are commonly applied to explore contrasting energy futures and to provide insights on how the energy system (or specific sub-sectors) may evolve under different conditions. However, ‘typical’ national level models are not fully adapted to capture the characteristics of local (city) transportation, and previous city-level ESOM based analyses have focused on decarbonisation of local energy systems, thus omitting other local policy considerations as e.g. air quality, and several studies excluded transportation altogether.  In this thesis, a generic city-level ESOM framework (TIMES-City) was further adapted and used to provide policy-relevant insights on the anticipated energy transition of the local transport sector. The underlying work rests on a systems analysis approach, building on careful consideration of the overall system performance and boundaries, understanding of specific system characteristics, and challenges and opportunities facing local ‘system managers’; this has implications for model representation and for quantitative and qualitative modelling assumptions. Further, availability and quality of local transport, energy and emission data needed to calibrate models poses significant challenges, and considerable effort was also put on producing projections for future transport demand (a key model input), using lessons and input data from traditional transport demand models. These considerations were addressed in Paper I.  The model was then applied to two different cases (in Sweden) to explore potential conflicts and co-benefits between ambitious climate targets and deep cuts in APs (Paper II), and to assess the roles of local and regional governments in CO2 mitigation when also considering ambitious national-scale policies (Paper III). The results of Paper II indicate that substituting fossil fuels for biofuels in conventional vehicles is the least-cost decarbonisation pathway, however this produces only minor or even negative benefits to air quality. While, zero-emission vehicles cut all local tail-pipe emissions, but their total impact on climate change mitigation is determined by upstream impacts from the conversion and distribution of energy carriers. Thus, ensuring low levels of total CO2 and APs from transportation calls for re-coupling of the local and global responsibilities and motivations into comprehensive mitigation strategies. The results of Paper III indicate that current Swedish national mitigation measures will drive down CO2 emissions in transportation considerably, but biofuel availability and BEV (battery electric vehicles) costs are critical for the rate and extent of the transition, while locally and regionally determined measures to enable shifts (from car) to active travelling, public transportation and home-based work have a much more limited direct impact. Nonetheless, these measures, along with city investments in BEVs and charging infrastructure which pave the way also for residents and local businesses, can help to reduce overall energy intensity of the transport sector, thus slowing down growth in fuel demand and contribute to reaching ambitious climate targets with limited renewable resources (as e.g. biofuels). The two studies (Papers II and III) illustrate the usefulness of applying comprehensive ESOMs also at sub-national levels, providing insights on both global and local sustainability implications as well as deeper understanding of the roles of local and regional decision-makers in enabling and supporting low-carbon transitions in transportation.

Objektų klasterizavimas – tai viena iš duomenų gavybos (angl. data mining) sričių. Šių algoritmų pagrindinis privalumas – gebėjimas atpažinti grupavimo struktūrą be jokios išankstinės informacijos. Magistriniame darbe yra pristatomas vaizdų klasterizavimo algoritmas, naudojantis savaime susitvarkančius neuroninius tinklus (angl. Self-Organizing Map). Darbe analizuojami vaizdų apdorojimo, ypatingųjų taškų radimo bei palyginimo metodai. Nustatyta, kad SIFT (angl. Scale Invariant Feature Transform) ypatingųjų taškų radimas bei aprašymas veikia patikimiausiai, todėl būtent SIFT taškiniai požymiai yra naudojami klasterizavime. Darbe taip pat analizuojamas atstumo tarp paveikslėlių radimo algoritmas, tiriami skirtingi jo parametrai. Algoritmų palyginimui yra naudojamos ROC (angl. Receiver Operating Characteristic) kreivės ir EER (angl. Equal Error Rate) rodiklis. Vaizdų klasterizavimui yra naudojamas ESOM (Emergent Self-Organizing Map) neuroninis tinklas, jis vizualizuojamas U-Matrix (angl. Unified distance Matrix) pagalba ir tinklo neuronai skirstomi į klasterius vandenskyros algoritmu su skirtingu aukščio parinkimu. Magistriniame darbe demonstruojami klasterizavimo rezultatai su pavyzdinėmis paveikslėlių duomenų bazėmis bei realiais gyvenimiškais vaizdais. / Clustering algorithms – a field of data mining – aims at finding a grouping structure in the input data without any a-priori information. The master thesis is dedicated for image processing and clustering algorithms. There are point-feature detection, description and comparison methods analyzed in this paper. The SIFT (Scale Invariant Feature Transform) by D. Lowe has been shown to behave better than the other ones; hence it has been used for image to image distance calculation and undirectly in clustering phase. Finding distances between images is not a trivial task and it also has been analysed in this thesis. Several methods have been compared using ROC (Receiver Operating Curve) and EER measurements. Image clustering process is described as: (1) training of ESOM (Emergent Self-Organizing Map), (2) its visualization in U-Matrix, (3) neuron clustering using waterflood algorithm, and (4) image grouping according to their best-matching unit neurons. The paper demonstrates the image clustering algorithm on public object image databases and real life images from the Internet as well.