Road Longitudinal Section Of The Algorithm Based On Ant Colony Optimization Design

The Optimization of Road Route provided the foundation for Automation of route selection by computer, and it is obviously superior to the plan comparison which is frequently used during road design. On the basis of plane alignment, Profile Optimization Design uses mathematic method to optimize the profile alignment, and possesses important meaning in reducing the project investment and improving the quality of route alignment design.This dissertation put forward a new optimization method for Profile Optimization Design through applying the ant colony algorithm. The chief research contents and achievements could be summarized as follows:(1) Based on studying the profile optimization design problem, Author set up the Mathematical Model of this dissertation which has three main aspects: choosing the design variable and the object function, setting up the constraint condition based on the code of road design.(2) By adopting mathematic technology to make a smooth curve which incarnates the basic landform rising and falling and changes the smooth area to make several different curves, the points of intersection of these curves intersect at the inflection points which show the switch of landform rising and falling. Using Minimum Two Multiplication Principle fitting beelines, every two beelines intersect at a point, and these points could be regarded as grade change points of the profile. Adjusting these grade change points based on Constraints to get the initial profile project.(3) Putting forward an optimization method for Profile Optimization Design based on Ant Colony Algorithm. Author divided the area around each grade change point into meshes in the initial profile project, set initial pheromone between every two central points of meshes, and simulated artificial ants by computer. The artificial ants can choose path according to choice probability and then update the pheromone. Authors used the artificial ants to optimize the positions of grade change points in the divided meshes by many times of iteration. At last, authors reduced the area and encrypted the meshes, repeated the process until the result is satisfactory.(4) Reached the conclusion that the ant colony algorithm needs suited parameters in order to get the best profile alignment by theoretical analysis and experimental verification.(5) Based on the theory expounded above, author developed The Profile Optimization Design Program using Visual C++6.0 and Object arx2002. The program has been verified by practices.