We design temporal description logics (TDLs) suitable for reasoning about temporal conceptual data models and investigate their computational complexity. Our formalisms are based on DL-Lite logics with three types of concept inclusions (ranging from atomic concept inclusions and disjointness to the full Booleans), as well as cardinality constraints and role inclusions. The logics are interpreted over the Cartesian products of object domains and the flow of time (ℤ, <), satisfying the constant domain assumption. Concept and role inclusions of the TBox hold at all moments of time (globally), and data assertions of the ABox hold at specified moments of time. To express temporal constraints of conceptual data models, the languages are equipped with flexible and rigid roles, standard future and past temporal operators on concepts, and operators “always” and “sometime” on roles. The most expressive of our TDLs (which can capture lifespan cardinalities and either qualitative or quantitative evolution constraints) turns out to be undecidable. However, by omitting some of the temporal operators on concepts/roles or by restricting the form of concept inclusions, we construct logics whose complexity ranges between NLogSpace and PSpace. These positive results are obtained by reduction to various clausal fragments of propositional temporal logic, which opens a way to employ propositional or first-order temporal provers for reasoning about temporal data models.
В рамках математического аппарата вероятностных нейронных сетей рассмотрена задача распознавания объектов по видеоизображению. Исследованы ее решения на основе построения коллектива решающих правил. Предложена оценка апостериорной вероятности принадлежности объекта к заданному классу для произвольного расстояния и метода ближайшего соседа. Показано, что эта оценка эквивалентна оптимальной наивной байесовской оценке, полученной для вероятностной нейронной сети с прямоугольной кернел-функцией, если в качестве меры близости используется информационное рассогласование Кульбака-Лейблера. Представлена блок-схема системы распознавания объектов из видео потока с автоматической адаптацией подаваемого на вход комитета классификаторов списка изображений идентичных объектов. Приведены результаты ее экспериментального исследования в задаче распознавания лиц для распространенных баз данных (FERET, AT&T, Yale).
We design a decidable extension of the description logic SROIQ underlying the Web Ontology Language OWL 2. The new logic, called SR+OIQ, supports a controlled use of role axioms whose right-hand side may contain role chains or role unions. We give a tableau algorithm for checking concept satisfiability with respect to SR+OIQ ontologies and prove its soundness, completeness and termination.
An initial-boundary value problem for the 1D self-adjoint parabolic equation on the half-axis is solved. We study a broad family of two-level finite-difference schemes with two parameters related to averages both in time and space. Stability in two norms is proved by the energy method. Also discrete transparent boundary conditions are rigorously derived for schemes by applying the method of reproducing functions. Results of numerical experiments are included as well.
The problem of developing a decision support system for estimating a) the scale of incorporating available renewable sources of energy (such as solar and wind energy) in a part of a country’s electrical grid (called a regional electrical grid further in this paper), and b) the scale of storing electricity in this (regional) electrical grid to make these renewable sources of electric power competitive with traditional power generators (such as fossil-fuel and nuclear ones) and to reduce the cost of acquiring electricity from all the electric power generating facilities in the grid is considered. In the framework of this system, renewable sources of energy are viewed as electricity generating facilities under both existing and expected electricity prices, and the uncertainty of energy supply from them and the uncertainty of the grid customer demand for electricity during every 24 h are taken into account. A mathematical model underlying the system allows one to study the interaction of all the grid elements as a game with a finite (more than three) number of players on a polyhedron of connected player strategies (i.e., strategies that cannot be chosen by the players independently of each other) in a finite-dimensional space. It is shown that solving both parts of the problem under consideration is reducible to finding Nash equilibrium points in this game.
There is a clear idea of how you can observe the patterns of eye movements. This is a dynamic and dynamic cognitive companion using a Dynamic Bayesian Network (DBN). Understanding how people live. This model provides a pre-frontal area for the colliculus. Our approach is to begin with. This is an analysis of the observer's task. Secondly, it is a state cognitive state. Finally, we’ve been able to make a difference. This is the only factor that influences observers' saccadic behavior. While the inclusion of the model, This is a list of the results obtained in the paradigms. It is capable of a saccadic simulation in real time. We have shown that it has been closely coordinated with those of human observers. It is important to note that it is not a problem. There are three ways to add top-down knowledge and knowledge. First of all, we’ve given the information system available to the visual system. This matches the influential signals by Milas and Cohen Annu Rev Neurosci 24: 2001 ) imple external shif shif Second, our model is a generative and capable of those found in visual search. Third, our model generates relative saccadic vector information as opposed to absolute spatial coordinates. This matches more closely together in the superior colliculus.