NETWORK METHOD OF EVOLUTION SYNTHESIS OF BEHAVIOR TESTS
Abstract
In this paper, we present a network decomposition method for synthesizing structured behavioral tests for use as part of diagnosis of distributed information systems (DIS). The network decomposition
method is based on the network evolutionary model of DIS testing, which contains a system of test, infecting,
realizing, transporting evolutions. The interaction of evolution takes place at three levels: within the component network for an individual network automaton; cooperatively component in the evolutionary implementation and transport of test behavior of a separate component of the automaton network representing DIS;
between combining component cooperations. The properties of the network interactions of the model made it
possible to determine the features of intercomponent basic automata compositions, parallel, sequential,particular, are the adaptability of mutations, that are altered by means of built-in infectious evolutions, the
preparation and repeated immune use of test, realized and transported behavior in the nodes of the automata
network, representing DIS. The definition of an algebraic system from base sets of test, realized and transported fragments, as well as signatures of compositional and evolution operations in the automata network,
made it possible to propose the basic structure of the network evolutionary method of decomposition synthesis of structured behavioral tests. The parallelism of test evolution, realizing and transporting evolutions of
the network model of testing is inherited by the network evolutionary method and allows parallel programalgorithmic realization of the interaction of evolution. As shown by analytical estimates and physical experiments, the network evolutionary method allows to reduce the time of synthesis of tests and preparation of
test support in general for the majority of real RIS in comparison with network deterministic methods.
feedback, evolutionary mutation and crossover operations, fitness and choice functions. Among them, in