METHOD FOR EXPRESS ESTIMATION OF FUNCTIONING OF COMPLICATED OPEN SYSTEM Russian patent published in 2007 - IPC G06N3/00 

FIELD: cybernetics, possible use for evaluating functioning of various systems, including complicated open systems.

SUBSTANCE: method is based on performing internal hierarchical formalization of complicated open systems with selection of factors which influence characteristics being researched (functions, properties, parameters) of complicated open system under scope, selection as main coefficients of functioning of complicated open system of complicated open system functioning level, which may be absolute and basic, setting of appropriate neuron network on basis of completed hierarchical formalization of complicated open systems, in which network neuron-like elements, in particular, neurons, model individual factors and system characteristic being researched, while a certain layer of neurons models factors of certain hierarchy level, computation of complicated open system functioning levels, after which numeric determined or expert estimate of condition and weight of factor of each input neuron-like element is set and normalized weight of factor of each neuron-like element is determined by division of signal, carrying information about weight value of current factor, by signal, carrying information about total value of weights of factors of neuron-like element, connected in synaptic fashion to appropriate neuron-like element of next level of hierarchy, factor parameter is determined for each factor of input neuron-like element by multiplication of signal, carrying information about value of estimate of factor condition, and signal, carrying information about normalized weight of factor of appropriate input neuron-like element, computed further is condition of factor of each neuron-like element of next hierarchy level, after that combined are signals, carrying information about parameters of factors of neuron-like elements, connected in synaptic fashion to appropriate neuron-like element of this hierarchy level, set next is numeric determined or expert estimate of weight of factor of each neuron-like element of this hierarchy level and analogically factor parameters of each neuron-like element of this level are computed, computed analogically is also condition and parameter of factor of each neuron-like element of following hierarchy levels, resulting state of factor of output neuron-like element represents level of functioning of complicated open system.

EFFECT: possible express evaluation of functioning of complicated open systems and, respectively, other, simpler systems, with appropriate qualitative and quantitative functioning estimates.

2 dwg