Расширение сетевых моделей на основе интеграции стохастических и нечетких графов

Abstract

Рассмотрены расширения сетевых моделей на основе интеграции стохастических моделей и нечетких графов. Показано, что создание гибридных моделей интегрирует достоинства частных моделей и исключает из рассмотрения недостатки этих моделей. Исследованы стохастические и нечеткие графы, которые являются модификацией графической модели оценки и пересмотра планов (ГЕРТ) и if/then-моделей. Исследование моделей показало достоинства и недостатки моделей, эффективность подходов подтверждена экспериментально в географических информационных системах.The paper solves the problem of expansion of network models based on the integration of stochastic models and fuzzy graphs. The object of the work is the processes of state estimation of spatially distributed objects. Scope of the work is expanding network models based on the integration of network and fuzzy graphs. The aim is to decrease the level of uncertainty and improving the efficiency of software applications based on the process modeling new hybrid models. It is proposed a hybrid model in the form of fuzzy Petri nets which is based on stochastic graphs. This model al-lows reducing the level of uncertainty without a significant increase in computational complexity. It is proved that the creation of hybrid models integrates the advantages of partial models and excludes from consideration the shortcomings of these models. It is studied the stochastic and fuzzy graphs. Stochastic graphs are a modification of the graphical model of evaluation and revision of plans (GERT), fuzzy graphs are extensions of if / then-models. It defines a set of restrictions on the use of stochastic and fuzzy graphs. It has been done the experimental research models which showed the advantages and disadvantages of models. Effective approach is confirmed experimentally in geographical information systems. Prospects for further research is the adaptation of structures to a wide class of problems on the set of alternatives. The works on implementation of software applications.