Application Of Ant Colony Algorithm For Geotechnical Back Analysis

Displacement back analysis was a well applied divaricator of Geotechnical Engineering developed in these years. It was based on optimization theory in order to estimate parameters of theoretical model by measurements of geotechnical excavation project. That means we can use the gray parameters to get a result in accordance with measurements to give a better solution of complicated geotechnical engineering problems.This thesis contains three parts, displacement back analysis based on traditional optimization theory, displacement back analysis based on Ant Colony Algorithm and the comparison between them. Traditional optimization theory finds its optimal solution by iteration of a single initial value. Because of its limitation, it always goes wrong to local optimal solution. Ant Colony Algorithm makes up the deficiency and it is based on positive feedback theory, and has characteristics of high robustness, distributed computing, manual heuristic information and apt to combine with other arithmetic, etc. Positive feedback theory could make the algorithm find its optimal solution faster, distributed computing and manual heuristic information could even help to find a better solution in initial stages and avoid premature convergence. In the thesis we have studied the numeric example and its application in project to get the geotechnical parameters and forecast displacement. We find that ant colony algorithm is better optimized compared with traditional optimization theory. The study shows that ant colony algorithm is global convergence in the combinatorial optimization and could avoid convergence prematurely.