Ant Colony Optimization And Its Improvement

The scientists study group-behavior characteristics of insects discovered that insects in community-level cooperation on the essentially is self-organization.In many occasions, although such cooperation may be very simple, but they can help to resolve complex issues, Take a colony of ants for example simple and blind ants can find the shortest routing path from the irnest to food source. The gregarious nature of this organism out of a collective act, that is generated by swarm intelligence has aroused many researchers interesting. Biologists had studied the Phenomenon carefully and found that ants cooperate to find the shortest routing path by means of indirect communications using a kind of substance call"pheromone".Inspired by this phenomenon a population based simulated evolutionary algorithm called ant colony algorithm was proposed by Italian researchers M.Dorigo,V.Maniezzo and A.Colorni in 1992.Many seholars areattraeted to study ant colony algorithm and in the past ten years than more the algorithm has been widely applied to the fields of combinatorial optimization,function optimization,system identification,network routing,path planning of robot,data mining and premises distribution of large scale integrated circuit etc,and good effects of applieation are gained.This paper focuses on the prineiples,improvement strategy,basic improvement algorithm and applieations of ACA, especially,in deep study on how to improve the basic ACO algorithm, inhibiting standstill of the algorithm, at the same time,test the improving algorithm by TSP. The main contributions of this dissertation are as follows:1,Propose an improved ant colony algorithm. It uses the moving of pseudorandom proportional,local pheromone update and global pheromone update method combining,we also improve the updated strategy,firstly,using regressive local pheromone update strategy,secongly introducting a new update strategy in global pheromone update.The improved algorithm expand the search space of solutions and avoid the algorithm into a local optimum.2,Propose a phased pheromone update ant colony algorithm. In early time of search ,the ants use the rapid accumulation of pheromone update strategy to achieve rapid decline in the algorithm;at the same tme we introduct the new update method which pull in the random variables factors. the new method makes the search for the optimal solution to update the pheromone on the degree of reduction,thereby reducing the gap of the pheromone on the path,which help ants search range become more broad,and succeed in finding the global optimal solution late in the search.Experimental results show that those two improved algorithms have more excellent exhibition in avoiding early variety,searching better solution and stability by simulation experiment and comparing to ant algorithm and ant colony algorithm.Finally, the work of this dissertation is summarized and the prospective of future research is discussed.