卷 45, 编号 5 (2018)

This paper presents a discrete model of non-synaptic interactions between neurons, which are performed by special chemical agents, neurotransmitters. In this model, all signals are broadcast: each neuron’s activity can be perceived by all other neurons that are sensitive to its neurotransmitter. The purpose of leaving synapses out is to demonstrate the importance of heterochemical¹ interactions in the nervous system. Transmitter-specific neurons are described as finite automatons, which are a generalization of the traditional formal neuron. In a network consisting of such neurons, all connections are functional and not structural. This feature makes it possible to explain rapid reconfigurations of activity patterns in the generation of different rhythms by neuron ensembles, which are observed in living systems. We discuss and justify the choice of a discrete mathematical framework for simulation of heterochemical interactions.

In this work, methods of formalization of the Unit–Function–Object system approach and the system–object method of representation of knowledge based on this approach are considered. The expediency of applying some ideas of the object calculation of Abadi–Cardeli and the pattern theory of Grenander for further formalization is substantiated. By analogy with the above algebraic apparatus, the calculus of special objects that represent elements of system–object models, which includes graphic formulism and basic operations with objects, has been developed. It is shown that using the proposed formal-semantic alphabet of special objects, it is possible to simplify the procedure for decomposition of a complex system on the basis of the basic hierarchy of classes of systemic connections. A substantiation of a number of system-wide regularities is given.

The concepts of the similarity and dissimilarity (difference) of analyzed objects play an important role in many theoretical and practical problems of decision making, artificial intelligence, pattern recognition, processing of heterogeneous information, etc. The similarity or dissimilarity of objects is usually estimated by their proximity in the attribute space. This paper considers new classes of metric spaces of finite, bounded, and measurable sets and multisets. The possibilities are shown for using new types of metrics to evaluate the similarity or dissimilarity of multi-attribute objects that are present in several instances with differing values of attributes and are represented by multisets of attributes.

Modern decision support systems (DSSs) use a broad range of technologies based on data processing by both humans and computers. This article describes a new architectural approach to creating DSSs based on the human–machine cloud concept. This concept combines two different views of decision support in the tourism industry: the standpoint of a tourist and the position of administrative bodies of tourist regions. The article distinguishes standard tasks of decision support in tourism; these tasks are then projected onto a multilevel concept architecture of cloud services. The proposed approach is illustrated based on the example of three scenarios of interaction of services (resources) in solving practical tasks.

This article presents the results of research conducted on cluster systems with multicore nodes to create a method and an algorithm that implement parallel inference in the proof of formulas in first-order logic. The algorithm is based on of J. Robinson’s resolution method of inference. A procedure for its parallelization, which allows one to dynamically control the uniform division of the set of resolvents distributed to the nodes of a computer system for inferring, is proposed. A number of heuristics, which make it possible to significantly reduce the time of implementing inference control decisions, are also introduced. Experiments here show that the effect on acceleration is higher than that of the previously studied parallel-inference algorithms.