In this manuscript we perform a rigorous mathematical investigation of the behavior opportunistic network models exhibit when two major real-world problems are taken into account. The first problem considered is obstruction. Here we model the network using an obstructed Gilbert graph which is a classical Gilbert graph but where there exist zones where no nodes are allowed to be placed. We take a look at percolation properties of this model, that is we investigate random graph configurations for which a component of infinite size has strictly positive probability to be created. The second problem considered in this thesis is mobility. Of course mobility in and of itself is not a problem but a feature in any network that follows the store-carry-forward paradigm. However it can be problematic to properly handle in a mathematical model. In the past this has been done by modelling movement by a series of static network configurations. However, with this technique it can be difficult to get a grasp on some of the time sensitive properties of the network. In this work we introduce the time bounded cylinder model which enables an analysis over a complete timeframe. We provide normal approximations for important properties of the model, like its covered volume and the number of isolated nodes. As we are using rigorous mathematics to tackle problems which computer scientists working in the field of distributed systems are faced with, we bring the two fields closer together.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第19112号 / 情博第558号 / 新制||情||99(附属図書館) / 32063 / 京都大学大学院情報学研究科知能情報学専攻 / (主査)教授 阿久津 達也, 教授 岡部 寿男, 教授 鹿島 久嗣 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

No description available.

No description available.

Artificial neural networks (ANN) have found extensive use in solving real-world problems in recent years, where their exceptional information processing is the main advantage. Facing increasingly complex problems, there is a need to improve their information processing. In this thesis, we explore new ways of synthesizing ANNs by reducing the synthesis problem to the Boolean satisfiability problem (SAT) that is, the problem of determining whether a given Boolean formula is satisfiable. Also known as the SAT problem, it aims to determine if there exists such a combination of Boolean variables in a propositional formula for which the formula evaluates to true. We derived a general formula in conjunctive normal form (CNF) representing the synthesis of a neural network. Given randomly generated datasets, we were able to construct CNF formulas whose satisfying assignments encode neural networks consistent with the datasets. These formulas were run through an off-the-shelf SAT solver, where the outputted models simulated the synthesis of neural networks consistent with the datasets. The experiments conducted in this thesis showed that our method had the ability to produce feed-forward neural networks of varying sizes consistent with randomly generated datasets of binary strings.

A de Bruijn cycle commonly referred to as a universal cycle (u-cycle), is a complete and compact listing of a collection of combinatorial objects. In this paper, we show the power of de Bruijn's original theorem, namely that the cycles bearing his name exist for n-letter words on a k-letter alphabet for all values of k,n, to prove that we can create de Bruijn cycles for multi-sets using natural encodings and M-Lipschitz n-letter words and the assignment of elements of [n]={1,2,...,n} to the sets in any labeled subposet of the Boolean lattice; de Bruijn's theorem corresponds to the case when the subposet in question consists of a single ground element. In this paper, we also show that de Bruijn's cycles exist for words with weight between s and t, where these parameters are suitably restricted.

Research in sustainable macroalgae aquaculture has been gaining hype (e.g. Seafarm) to provide advancements in new significant sources of food, feed, materials and bio-energy. Despite the fact that in Asia the offshore production of macroalgae is established for many decades, in Europe is still in its infancy. The issue on where to find suitable cultivation areas without conflicting with current uses to respect the environment and the socio-economic activities is a great challenge. This Master of Science thesis aimed to find suitable offshore areas in order to facilitate implementations of macroalgae cultivations in the Swedish West Coast. Thirteen criteria in environmental, economic and social sustainability aspects were identified and employed (e.g. Depth, Distance to Ports, and Natural and Preserved areas (NPAs)), and the tools Geographic Information Systems (GIS) and Multi-criteria Analysis (MCA) were used in the form of GIS-MCDA, an integrated method available in Idrisi. To aggregate the criteria, Boolean and Weighted Linear Combination (WLC) techniques were applied. The results showed that Boolean areas cover 537 km2 (6,98% of the study area). The best suitable areas, possessing the maximum suitability index 10, that resulted from two WLC models comprise 5 km2 (0,07% of the study area) and 26 km2 (0,34% of the study area) including and excluding the criterion NPAs as constraint, respectively. The results further indicated that GIS-MCDA models excelled in providing an overview for effective spatial decision-making. Both techniques play a role in suitability analysis and complement each other in finding an optimal site which could be carefully selected out of the identified areas. It is recommended that areas with a suitability index 10 be chosen inside Boolean suitable areas. Moreover, this study could act as a driving force to build a resilient planning framework that would boost sustainable placement and development of offshore macroalgae cultivations.

No description available.

This project is about finding ways to improve performance of a Computer-Aided-Design (CAD) application running in the web browser. With the new Web API WebGPU, it is now possible to use the GPU to accelerate calculations for CAD applications in the web. In this project, we tried to find if using the GPU could yield significant performance improvements and if they are worth implementing. Typical tasks for a CAD application are split and union, used for finding intersections and combining shapes in geometry, which we parallelized during this project. Our final implementation utilizes lazy evaluation and the HistoPyramid data structure, to compete with a state-of-the-art line-sweep based algorithm called Polygon Clipping. Although the Polygon Clipping intersection is still faster than our implementations in most cases, we found that WebGPU can still give significant performance boosts.

No description available.