Vol 44, No 5 (2017)

This review addresses the issues of computations using cellular automata (CA). It is shown that the generality of the connectionism paradigm allows some methods applicable to neural networks to be transferred into the domain of CA. Some special issues of computability are discussed based on the examples of the density classification task, the firing-squad synchronization problem, and the queen-bee problem, as well as sorting algorithms and Atrubin’s parallel multiplication algorithm.

This article reviews a new variant of the generalized JSM-method for automatic support of scientific research (ASRS). This version of the JSM-method applied for ASRS contains ternary predicates of similarity (cause, brake, effect) which contain binary “cause–effect” predicates and explicit brake definitions. The article considers a possibility for the basis of fact partition for the cases of the truth of the generalized method and an inductive method of similarity with a ban on counter-example predicates.

In this paper, we consider some game-theory problems of group pursuit of a target under perturbations. Here, the players are unmanned flight vehicles (FVs) whose mathematical models are given by transfer functions that describe a double-loop control system with an autopilot and certain settings to provide the necessary stability of flight. According to the separation principle, without loss of generality, solutions are considered in a pitch plane. In the case of an antagonistic game, the velocity of the target is higher than that of the pursuers. The problem is solved when one of the pursuers gets close enough to the target or when the target manages to evade the pursuers. The target tracking problem implies that a randomly-arranged FV group approaches the target and flights near it during a specified time period. The low-velocity target seeks to evade the pursuers as far as possible. Finally, in the path following problem, each FV needs to fly along the trajectory given by the motion of a corresponding reference target. In the process of problem solving, each FV implements a set of heuristic behavioral strategies in a perturbed environment by following the rules of pitch angle and velocity selection. In the experimental part of this paper, some situations typical for these problems are modeled.

In this paper, the multicriteria choice problem, including the decision maker’s (DM) binary preference relation, is considered. Two combined two-stage approaches are proposed. According to these approaches, first, we have to reveal some information on the DM’s preference relation and then apply scalarization methods based on linear or multiplicative functions. The justification of both the combined two-stage approaches is given.