The Improvement And Application Of Artificial Bee Colony Algorithm

Artificial Bee Colony (ABC), proposed by Turkey scholar D. Karaboge in2005,is a kind of swarm intelligence optimization algorithm. The algorithm was favoredby scholars and experts from many countries in a very short time. It has beensuccessfully applied to the neural network, distributed computing network andmulti-objective optimization problems, and other fields, and has achievedoutstanding results. No one is perfect, so it is with this algorithm. Although the ABCalgorithm shows its unique advantages in many ways, there are some shortcomingsfound by scholars, such as the algorithm easily falls into local optimum, thealgorithm convergence speed is too slow in the late, operation process of algorithmhas nothing to do with its iteration times，not only need algorithm in the aspects oftheory and design improvement still needs to extend application scope.This paper makes some research of ABC pointed at above problems. Theresearch results can be summarized as follows:(1)An improved artificial colony algorithm with self interference is put forward.In this algorithm introduced a number of iterations and the algorithm itself about theinterference coefficientj, changed the algorithm relies on a random coefficient. Itmakes the algorithm of iterative associated with its number of iterations, lets notcontrol before iteration into regulation of iteration.jis a monotone decreasingfunction at the same time, this factor can not only guarantee the algorithm in theearly time to keep a wide range of exploration ability; but also can guarantee thealgorithm in the late to have a smaller range of mining capacity, with greatly reducesthe late original algorithm convergence speed too slowly.(2)On the original algorithm with roulette wheel selection strategy makesalgorithm easily fall into local optimum, this paper puts forward the improvedalgorithm based on the update mechanism. When the algorithm falls into localoptimum, by a party update mechanism and the introduction of new food source toincrease the diversity of the bee population, which the algorithm can quickly jumpout of local optimum.(3)ABC is an optimization algorithm, which the problem solved is to find the (4)optimal solution of the problem. Through that we can combine clusteringproblem solving with clustering problem of clustering center, completed the fusionof two kinds of algorithm, and achieved good results.