We address the task of mapping a given domain model (e.g., an industry-standard reference model) for a given domain (e.g., ERP), with the source code of an independently developed application in the same domain. This has applications in improving the understandability of an existing application, migrating it to a more flexible architecture, or integrating it with other related applications. We build on a previous approach, which uses relationships among source code elements for improving the precision of the mapping process. We extend this approach by considering relationships among domain model elements in addition to relationships among source code elements, and also by stating the mapping process as an optimization problem. We have implemented our approach, and compared it with the previous approach. We show that our approach gives significantly better precision as well as recall than the previous approach when applied on a real industry-standard domain model and an open-source application.

abstract: Current work in planning assumes that user preferences and/or domain dynamics are completely specified in advance, and aims to search for a single solution plan to satisfy these. In many real world scenarios, however, providing a complete specification of user preferences and domain dynamics becomes a time-consuming and error-prone task. More often than not, a user may provide no knowledge or at best partial knowledge of her preferences with respect to a desired plan. Similarly, a domain writer may only be able to determine certain parts, not all, of the model of some actions in a domain. Such modeling issues requires new concepts on what a solution should be, and novel techniques in solving the problem. When user preferences are incomplete, rather than presenting a single plan, the planner must instead provide a set of plans containing one or more plans that are similar to the one that the user prefers. This research first proposes the usage of different measures to capture the quality of such plan sets. These are domain-independent distance measures based on plan elements if no knowledge of the user preferences is given, or the Integrated Preference Function measure in case incomplete knowledge of such preferences is provided. It then investigates various heuristic approaches to generate plan sets in accordance with these measures, and presents empirical results demonstrating the promise of the methods. The second part of this research addresses planning problems with incomplete domain models, specifically those annotated with possible preconditions and effects of actions. It formalizes the notion of plan robustness capturing the probability of success for plans during execution. A method of assessing plan robustness based on the weighted model counting approach is proposed. Two approaches for synthesizing robust plans are introduced. The first one compiles the robust plan synthesis problems to the conformant probabilistic planning problems. The second approximates the robustness measure with lower and upper bounds, incorporating them into a stochastic local search for estimating distance heuristic to a goal state. The resulting planner outperforms a state-of-the-art planner that can handle incomplete domain models in both plan quality and planning time. / Dissertation/Thesis / Ph.D. Computer Science 2014

This thesis conceptualizes a generic monitoring framework for online markets, which has also been implemented in a prototypic fashion. Thereby identifying a set of arising challenges for which solutions have been developed. An introductory section gives a short overview of the field of research, states identified scientfic challenges and derives research questions thereof. The following articles describe (i) the general approach of an online market monitor, (ii) how to encapsulate domain-dependent configurations & functionalities from monitoring core modules to facilitate a generic approach, (iii) change frequency estimation for observational units in a dynamic and fuzzy population setting, (iv) the development of an adaptive harvest heuristic scheduling new observations by utilizing the change frequency estimator, (v) how to make use of collected market information in form of business intelligence reports and finally (vi) an exemplary meta-analysis showing how to draw further conclusions about market mechanisms.

Although the overall levels of harmonics in modern power supply systems are in most of the practical cases still below the prescribed tolerance limits and thresholds (e.g. these stipulated in [IEC 61000-3-2 and 61000-3-12]), the sources of harmonics are constantly increasing in numbers and are expected to increase even more in the future. Some of the examples of modern non-linear power electronic (PE) devices that are expected to be employed on a much wider scale in LV networks in the future include: light-emitting diode (LED) lamps, switched-mode power supplies (SMPS'), electric vehicle battery chargers (EVBCs) and photovoltaic inverters (PVIs), which are all analysed in this thesis. The thesis first reviews the conventional harmonic analysis methods, investigating their applicability to modern PE devices. After that, the two most widely used forms of harmonic models, i.e. component-based models (CBMs) and frequency-domain models (FDMs), are applied for modelling of the four abovementioned types of modern PE devices and their models are fully validated by measurements. The thesis next investigates the impact of supply voltage conditions and operating modes (e.g. low vs high operating powers) on the device characteristics and performance, using both measurements and developed CBMs and FDMs. The obtained results confirm that both supply conditions and operating modes have an impact on the characteristics of most of the considered PE devices, which is taken into account in the developed models and demonstrated on a number of case studies. As the next contribution, the thesis proposes new indices for the evaluation of current waveform distortions, allowing for a separate analysis of contributions of low and high frequency harmonics and interharmonics to the total waveform distortion of PE devices. As the modern PE devices are normally based on high-frequency switching converters or inverters, the impact of circuit topologies and control algorithms on their harmonic emission characteristics and performance is also investigated. Special attention is given to the operation of PE devices at low powers, when there is a significant increase of current waveform distortion, a substantial decrease of efficiency and power factors and when input ac current might lose its periodicity with the supply voltage frequency. This is analysed in detail for SMPS', resulting in the proposal of a new methodology ("operating cycle based method") for evaluating overall performance of PE devices across the entire range of operating powers. Finally, a novel and simple hybrid harmonic modelling technique, allowing for the use of both time-domain and frequency-domain models in the same simulation environment, is proposed and illustrated on the selected case studies. This is accompanied with a frequency-domain aggregation approach, which is applied in the thesis to investigate the impact of increasing numbers of different types of modern PE devices on the LV network. The implementation of the developed hybrid harmonic modelling approach and frequency-domain aggregation technique is demonstrated on the example of a typical (UK) urban generic LV distribution network and used for the analysis of different deployment levels of EVs and PVIs. The presented harmonic modelling framework for individual PE devices and, particularly, for their aggregate models, fills the gap in the existing literature on harmonic modelling and characterisation of modern PE devices, which is important for the correct evaluation of their harmonic interactions and analysis of the impact of their large-scale deployment on the overall network performance.