Adaptive Ant Colony Optimization Algorithm

Ant colony algorithm is an algorithm of simulating biological ants in the behavior of finding food source. The imitation of natural disposition and the strong local search ability are the characteristics of this algorithm. There are characteristics in solving optimization problems, which reflect distinct advantage. such as positive feedback and robustness characteristics, otherwise be combined with other bionic optimization algorithm easily.Ant colony algorithm has been used widely. such as the processing of the TSP problem, the transportation, the pipe laying and the plant selection, etc. However, the applications always meet difficulties, especially the problem of complexity.If you solve the problem with the basic ant colony algorithm, the algorithm will fall into local optimum and the chances for stagnation phenomenon will be greatly improved.In such way the precision and the convergence speed also cannot be guaranteed. For such problems, a lot of scholars and experts attempt to put forward many improved ant colony optimization algorithms, such as the elite strategy of ant colony algorithm, ant colony system, maximum minimum ant colony system and so on. Those improved ant colony algorithm in the solution of the optimal solution accuracy has improved significantly, but there are some problems to be solved, such as the longer search time in initial time and the no adaptability of volatile factors in the global update rules.For the above defects, the basic ant colony algorithm in this paper the initialization pheromone distribution and global pheromone volatilization factor do optimize respectively as following:Firstly, the initialization pheromone concentration distribution related to the distance.In other words,it has the direction at the initial moment.In such way,ant colony algorithm will speed up the initial search and avoid ant colony wasting more time blindly in the initial phase random search, which increase the high quality solutions.Secondly, global volatile factor has not adaptive in the process of global pheromone update, so to join the hyperbolic tangent function as its volatile dynamic factor.The purpose of this way is to make it adaptive to the refresh each iteration the optimal path pheromone concentration smoothly.In such way, it increases the likelihood of algorithm to obtain the global optimal solution and global search ability improves continuously, also the stagnation phenomenon in a certain range has been avoided.The experimental results show that the improved algorithm not only ensure convergence speed, increase the likelihood of algorithm to obtain the global optimal, but also is suitable for solving large-scale optimization problems.In this paper, two examples of shoals of ant with the improved algorithm is optimized and the optimized result with the actual have good consistency, illustrates the effectiveness and practicability of the improved algorithm.