The Research And Application Of Artificial Bee Colony Algorithm Based On Shared Factor

As a new meta-heuristic optimization algorithm, artificial bee colony(ABC) algorithm has been the concern of many experts and scholars. The algorithm was proposed by the Turkish scholar Karaboga in 2005, inspired by the foraging process of honey bees. For artificial bee colony algorithm, there are many advantages, such as fewer parameters, easy to implement, simple computation, fast convergence and so on. Therefore, it has been applied in many fields. However, ABC algorithm has the poor performance in solving function optimization problems, and is easy to fall into local optima in the process of algorithm implementation and has poor convergence rate. Therefore, this article has studied how to improve the performance of ABC search algorithm and apply it in the Xinanjiang hydrological model. The main contents of the thesis are:First, we detailed analyzed the biological principle, mathematical models, algorithms processes of standard ABC algorithm. According to the characteristics of the ABC algorithm,we choose several improved ABC algorithms from the related literature, and mainly study an improved ABC algorithm, namely, Gbest-guided artificial bee colony(GABC) algorithm that was presented by Zhu etc. The GABC algorithm increases the global search process of the ABC algorithm which is randomness and blindness in search solution. The GABC algorithm and the standard ABC algorithm were taken on the same set of benchmark functions. Test results show that the GABC algorithm is higher than the ABC in searching accuracy, and faster in convergence speed.Secondly, through the study we found that GABC algorithm does not make the balance of the global search ability and local search ability, and there is plenty of space to improve search precision and convergence speed. Therefore, this article improves the GABC algorithm's search equation and proposes global optimal artificial colony(SF-GABC)algorithm based on the shared factor. We added two share factors respectively in the search equation of local search and global search to make the honey bees searching purposefully rather than blind searching. Therefore, the optimization ability of the algorithm was enhanced.In order to testify the validity of the improved SF-GABC algorithm, under the same parameter settings, we made the ABC algorithm, GABC algorithm and SF-GABC algorithm simulation experiments on a set of standard test functions respectively, analyzed and compared the results of experiment. Experiment results show that SF-GABC algorithm convergence speed is faster, and function optimization precision is higher.Finally, the improved algorithm was applied to the XinAnJiang hydrology model to solve the optimization problem of parameter estimation. And we analyzed and compared the parameter estimation results of ABC algorithm, GABC algorithm, SF-GABC algorithm. Theexperiment results show that the model parameters' optimization accuracy of SF-GABC algorithm is higher and the convergence speed is faster than the ABC algorithm, GABC algorithm.