Abstract: It is a key problem that how to determine the most dangerous sliding circular surface and calculate the minimum safety factor in the stability analysis of soil slope. The traditional limiting equilibrium analysis method will be easily immersed in local minimum, so the real most dangerous sliding circle can not be searched out. For the reason, a new approach which is genetic algorithms integrating the traditional method of slices with vertical interface is put forward to analyze the entire stability of soil slope. The stability analysis model of soil slope which can be called object function is established by the methods of slices with vertical interface; and then the most dangerous sliding circular surface will be searched out by the genetic algorithm. So the minimum safety factor of soil slope will be figured out. The analysis process will be performed by the computer program designed by the Microsoft Visual C++ program language. After the search scope of the most dangerous sliding circular arc is determined and input into the program, all the calculation and analysis work will be done by the program algorithm automatically. Genetic algorithm simulates genetic evolutionary process of organisms and avoids the limit of traditional methods successfully. Compared with the minimum safety factor figured out by the area subdivision method and the most dangerous sliding circular arc searched out by the same means, the conclusion can be made that the stability analysis result of soil slope analyzed by the genetic algorithm is much more reliable. Genetic algorithm can successfully solve the problem that how to search the entire extremum of soil slope. The engineering application case indicates that the genetic algorithm is effective to analyze the stability of soil slope, which has extensive prospect and will be widely used in the slope engineering.

Key words: genetic algorithm, minimum safety factor, soil slope, entire stability.