Research On The Ant Colony Optimization Algorithm And Its Applications

As a typical example of the animals living in groups, an ant in the colony can produce self-organization behavior in coordination and cooperation, meanwhile its individual behavior is single. For example, an ant colony can always find out the shortest path between their cave and the food. Inspired by that, an Italian scholar, M. Dorigo, together with his group put forward an algorithm called ant colony algorithm. They simulated the foraging behavior of ant colonies and worked out the ant colony algorithm through mathematical modeling and experiment research, which was later applied for solving traveling salesman problems and achieved good results. With the continuous efforts of scholars to study and improve it, the ant colony algorithm has gradually become an independent branch of the artificial intelligence algorithms in recent twenty years. And now it has been widely used in various fields, like system identification, network routing, portfolio optimization, path planning, data mining, image processing, integrated circuit layout design, and function optimization.This paper systematically studied the principles, the models, and the applications of ant colony algorithm. And it has an especially deep study of the solutions to the improvement of basic ant colony algorithm and its applications for the field of noise image edge detection as well. In this paper, an improved ant colony algorithm in the field of earthquake fault tracing application was also studied. The main achievements of this paper include: 1. Based on AC model, a large number of simulation experiments were did for the determination of the important parameters for the AS algorithm. And it has obtained the optimal value of parameter selection following the analysis of experimental data.2. An exponential model of ant life cycle strategy was put forward aiming at the defects of slow convergence of ant colony algorithm. It set the first cycle ant path length obtained as the life cycle of standard-setting ant, thus successfully weakening the influence of the poor solution on the subsequent runs of the algorithm by enhancing its positive feedback feature.3. The probability selection function of ant colony algorithm, pheromone update strategy and heuristic factors are improved in solution to the problem that the algorithm is easy to fall into stagnation. Compared with the original algorithm, its Inhibition of stagnation of the algorithm plays a certain effect.4. The noise image edge detection for this specific application of ant colony algorithm has been improved, which includes the selection of the initial position of ants, adding enlightening factor to the image edge features and so on. Improved algorithm is applied to the noise image edge detection, the right edge of the image noise suppression and achieved good results.5. The improved ant colony algorithm is applied to the field of earthquake fault trace, using ant algorithm to detect the edges of tomographic images, which enhances fault discontinuity, and lays the foundation for subsequent processing.