Research On Improved Ant Colony Optimization And Its Application In TSP

Since the concept of swarm intelligence is put forward, lots of corresponding intelligent methods and intelligent algorithms are also put forward which are being used in solving problems that can't be dealt with by classic algorithms. According to the actual applications and the results of experiments, many intelligent algorithms could solve some specific problems very well as well as the good result. At the meantime, as a representative group of intelligent algorithms the ant colony algorithm had become a hot spot in the research.Started by the strategy of ant foraging process, it formalizing modeled the ant foraging and performed the quantitative relationship between the various parameters and the system behavior via solving the model in this paper. Then according to the different distribution of food sources, this experiment analized the influence which the food source imposed on group behavior. Through the analysis of the number of group and the evaporation rate of the pheromone, it can show that there is a power index relationship between the population size and the performance of the system. It showed that the increase of the group population contributed a lot to the performance of the system and proved this kind of quantitative relationship can be directly used to both theoretical and experimental support for the ant colony optimization algorithm in this paper. It uses TSP as an example to do a deep research based on the above mentioned conclusion in this paper. First, it introduced rules and formal descriptions of the ant colony algorithm for TSP. Then according to the result of the analysis given by the ant colony foraging model, the ant colony algorithm is improved in this paper. And based on the experiment result of the evaporation effects and the parameter of the population size, an ant colony algorithm-improving algorithm which introduced a memory mechanism and the transmissions of ant is designed in this paper.This thesis also implemented certain experiments for the formalizing models and improved ant colony algorithm to prove the correctness of the theories and the improvements.