Приложение теории искусственного интеллекта в задачах моделирования устойчивости грунтового массива

Abstract

Paper is devoted to a problem of ensuring stability of the soil massif by use of an automatic control system. Today the main problem of the soil-cultivating units used on slopes and also in the conditions of a difficult relief and existence of obstacles is uncontrollable change of a trajectory of their movement. It is possible to solve this problem by introduction of the computer control and management systems having possibility of fast adaptation in the course of work. The carried-out analysis of literature showed that now there are no computer monitoring systems of stability of the soil massif (landslide) that would allow units to react quickly to extreme conditions and to adapt for them. Authors offer the concept of a control system, the main on creation of automatic indistinct system. This approach is realized by means of a program complex, data of supervision move on an entrance of system which makes the decision on danger of emergence of instability. Thus this system is adjusted optimum for obtaining more exact decision by means of genetic algorithms. Creation of model of forecasting of stability of the soil massif (embankment) is considered, that is the necessary condition of emergence of a landslide is considered. For the forecast of the coefficient influencing stabil-ity methods of fuzzy logic are offered. We construct linguistic variable that takes some value for the terms of the membership function is constructed, built a knowledge base, fuzzification and defuzzification is performed, obtained some prognostic value that specified by the genetic algorithm. The algorithm of emergence of instability pays off and fast response of the operator to this phenomenon is predicted, that is the sequence of the actions necessary in order to avoid danger is specified.The development of stability ground mass worked out. The fuzzy logic methods are considered for predication of development stability ground mass.