Evaluation of Bank Slopes Stability Based on Random Ants Clustering Algorithm

Abstract

There are many factors affect the stability of bank slopes, each of them is associated and coupled with others. The analysis of slopes stability can be achieved by the method of effect-factors analogy and cluster analysis. Traditional cluster analysis is difficult to obtain the stable global optimal solution, since the results are sensitive to the initial cluster center and the order of sample input. Conventional ants clustering algorithm can get an appropriate result, by simulating the ants’ intelligent behavior of transportation. Consequently, in this paper a method of ants clustering algorithm base on random disturbance is proposed which can gain the cluster analysis of bank slopes through guiding the ants cluster by the accumulation and change of ant pheromones. The results, compared with conventional cluster analysis method, simulated annealing algorithm method and conventional ant clustering algorithm, shows that this method is effective and convenient to accomplish the analysis of bank slopes.