Improvement And Research Based On Bee Evolutionary Genetic Algorithm

Search engine solves the problem of rapid information retrieval. However, as the search methods difference, information retrieval efficiency and accuracy will be different. Meta search engine composites the advantages of each search engine, through the scheduling of each search engine, to get better effect. The current scheduling algorithms cannot coordinate the relationship among recall, precision and response time well.Genetic algorithm is a kind of common intelligence algorithm with widely application. Bee evolutionary genetic algorithm is an improved genetic algorithm based on bee genetic regularity. Since the random population size influents on the performance of the bee evolutionary genetic algorithm, this thesis puts forward some improvement to this algorithm. And it was applied to the multi-dimensions function optimization and meta search engines scheduling, and then obtained good results. The main research results are as follows:(1)Through the analysis of the bee evolutionary genetic algorithm, an improved bee evolutionary genetic algorithm was proposed; the algorithm uses the strategy by stages to adjust the dynamic random population scale. The random population size changing gradually, guarantees not only the diversity of population, but also improvement of convergence speed and accuracy for the algorithm. The experimental results by the optimizations of typical high dimensional function show that the algorithm is effective and feasible.(2)The improved bee evolutionary genetic algorithm will be applied in the scheduling of meta search engine, as well as dynamical optimization of many independent search engine. When Multi-objective combinatorial optimization scheduling cannot get the most optimal at the same time, the fitness of each objective function will coordinate the combination of each search engine optimization by adjusting individual comprehensive fitness weight, finally find the search engine combination of non-inferior solution scheduling list. The experimental results show that the proposed algorithm enhances the scheduling efficiency and is better than conventional technique in precision and speed.