Recurrent Models of Randomized Processes in Foreign Intelligence Tasks. Part 2. Simplified High Order Models

Abstract

Foreign intelligence at all times has always remained the sphere of special attention of many states. Its significance has particularly increased at the present time in connection with the acute nature of their relationship. New forms of conflicts have appeared: network and hybrid wars, large-scale terrorist acts, the space for their conduct has become more complicated − cyberspace has been added to the traditional maritime, land and air spaces. The development of foreign intelligence tools is a long and expensive process, requiring constant additions and changes in accordance with events in the international arena. However, the use of the theory of random processes for these purposes encounters one very difficult problem. The fact is that the classical theory implies the processing of data under conditions of equidistant observations, while in reality intelligence data is a stream of data obtained at random times. Moreover, a random amount of data may arrive at random times and, finally, there may be a situation where there is so small quantity of intelligence data that it is impossible to draw any conclusions on the decision making. For such very often encountered conditions, spline models of varying degrees of complexity are proposed in the work, which allow synthesizing adaptive algorithms with precisely known and unknown moments of changes in the state of the reconnaissance object. The Poisson process is used as an event flow model. The correctness of the algorithms is ensured by the mathematical rigor of the above reasoning; the results of simulation are presented.