京都大学 / 0048 / 新制・課程博士 / 博士(農学) / 甲第22501号 / 農博第2405号 / 新制||農||1077(附属図書館) / 学位論文||R2||N5281(農学部図書室) / 京都大学大学院農学研究科森林科学専攻 / (主査)教授 柴田 昌三, 教授 神﨑 護, 教授 德地 直子 / 学位規則第4条第1項該当 / Doctor of Agricultural Science / Kyoto University / DFAM

The thesis is dealing with implementation of IMPRODEMO image processing application. Author has used existing image processing libraries and has implemented complex application with modern user interface. In order to develop the application, he used framework WPF, method for developing application using tests called Test Driven Development, design pattern MVVM and MEF tool for easy integrating of extensions. Chapter two deals with these technologies and methodologies and describes them. Third chapter describes design of application and presents class diagrams. Fourth chapter describes test driven development and particular pitfalls that emerged from using this methodology. Fifth chapter presents the mean of integrating code written in C++ language into .NET framework and into C# language. At the end, whole work is summarized, resuls are defined and possibilities of resuming and extending the work are proposed.

<p>Effects of prairie restoration on arthropod diversity was investigated at Gabis Arboretum, Valparaiso, Indiana. A total of 35,408 arthropods belonging to 13 taxa in the restored prairie (RP1 and RP2), old field (OF), and monoculture stand of Phalaris arundinacea (reed canary grass – RCG) sites, were captured, counted, and compared. The enhanced plant species diversity in the restored prairies did not appear to promote the diversity of arthropod taxa. However, the restoration led to a more balanced composition of arthropod functional groups and thus elevated the diversity of functional groups. The arthropod assemblages in the three sites diverged clearly according to my canonical correspondence analysis (CCA) ordination. Pollinator abundance was greatest at RP and least at RCG site, positively correlating with greater forb diversity, and suggesting greater potential for nectar feeding and pollination potential at RP sites. Herbivore abundance was greatest at the RP sites, positively correlating with increasing plant species diversity. Predator abundance was significantly greater at the RCG site compared to the OF and RP sites; it was positively correlated with greater C3 grass cover, a characteristic of the structurally homogenous RCG site, and negatively correlated with increasing plant diversity and forb cover, a characteristic of the diverse and more structurally complex RP sites. Given the apparent non-random distribution of arthropods among the field types, my results suggest plant species composition has a significant effect on arthropod assembly. The monoculture grass stand was found to have a predator dominated arthropod community supported by a small, diverse herbivore community. It is concluded that the prairie restoration has resulted in alteration of arthropod communities supporting greater pollinator and herbivore abundance and a more balanced ratio of herbivores to predators due, in part, to increased plant structural diversity.</p>

There has been almost fifteen years of political controversy surrounding changes to the rules and procedures by which aircraft conduct their flight within regional Australia. Decisions based on a predominately heuristic (rule of thumb) approach to design have had many adverse consequences for the integrity of the proximity warning function. A sound mathematical model is required to establish this function on a mature engineering foundation. To achieve this, the proximity warning function has been investigated as a hybrid-system. This approach recognises the dual nature of the design: that aircraft dynamics give rise to continuous mathematical models while the communication protocols controlling proximity require discrete mathematical approaches. The blending of each aspect has yielded a deeper insight into the operational limitations and failure modes of this function. The presentation of the thesis follows a design thread through the function. It begins with a description of existing standards and implementations. Risk models are then developed. The pilot interface is recognised as a primary design constraint. Mathematical models are then developed to describe the topology of flow, proximity dynamics, and the scheduling constraints associated with visual, voice, and data-link communications required by the proximity warning function. These analyses show that many aspects of design can be bounded by analytical formulae that bring new robustness to the design and resolve some of the misconceptions arising from the often inaccurate perceptions of present airspace operations. Failure modes, unaccounted for in existing designs are found to actually aggravate failure in the very situations in which the airspace design should be robust and should act to prevent collisions. In particular, there are divergences of performance between the demands required by the system design and the ability of the pilot to deliver such performances. In some cases, these failures may be traced to policy decisions such as service between Instrument Flight Rule and Visual Flight Rule category aircraft. On the basis of the conclusions of this research, a formal engineering review of the proximity warning function is required to assure the containment of the likelihood of mid-air collision for all future operations.