The Global Optimization Of Hing Dimension Multi Extremum Functionbased On Artificial BEE Colony Algorithm Abstract

Artificial Bee Colony algorithm, after more than a decade of research and application, has developed to a certain stage. However, like other random bionic optimization algorithm, when solving the problem of high-dimensional, algorithm is easy to fall into local optimal solution and slow convergence speed. To solve these problems, this paper will present some improved methods about artificial bee colony algorithm, then use the improved algorithm to solve high-dimensional complex functions of multi-optima global optimization problem, and make the corresponding analysis.The main research work in the paper is included as follow:(1) This paper, after analyzing the current commonly used several kinds of fitness function transform,put forward the fitness function, based on iterative algebra and dynamic change of the exponential transform, then use the fitness function to solve high-dimensional optimization problem of multi-optima function. The results show that the improvement of the fitness function in this paper is effective.(2) In order to optimize of individual evolutional iteration strategy, This paper proposes a searching strategy for a dynamic control lead bees or follow bees to search space the field of nectar source, namely, at the beginning of the algorithm search, the search space of the lead or follow bees for nectar source is much larger, and the probability of the individual species to find more and high quality new bee is bigger, Thus to reduce the probability that the algorithm trapped in local optimum, and increase the probability of finding the global optimal solution. Review at later time of the algorithm search, Appropriate to reduce the neighborhood searching space of the lead or follow bees, in order to improve the local search ability, the convergence speed and the precision of algorithm. The thesis apply the improved algorithm for global optimization to make a simulation experiment on multiple high dimensional and multipolar function, the results show that the effectiveness of the improved algorithm is proposed. In this paper, we use the improved algorithm to optimize the global optimization of several high-dimensional multi-extremum functions. The results show that the proposed algorithm is effective.