Improved Flower Pollination Algorithm And Its Application In User Identification Across Social Networks

With the increasing scale,larger dimension and higher complexity of various optimization problems,traditional optimization techniques can no longer meet the growing demands.The proposed meta-heuristic algorithm provides a new research direction in the field of optimization.The flower pollination algorithm is one of them—a new kind of meta-heuristic intelligent algorithm,the idea of which is derived from the process of plant flower pollination.The flower pollination algorithm has few parameters,strong searching ability and high robustness,which has been widely used in the optimization problem of various fields.However,the pollination algorithm itself has the disadvantages of slow convergence and easy to fall into local optimum at a later stage.In view of the shortcomings of the flower pollination algorithm,we make further analysis and improvement,and apply the improved flower pollination algorithm to the user identification problem in social networks.The main work includes the following aspects:(1)A flower pollination algorithm(DCFPA)based on dynamic global search and Cauchy mutation is proposed.The algorithm uses chaotic map to enhance the randomness and uniformity of the initial distribution of pollen population.The average optimal pollen position and dynamic weight reduction factor are introduced based on the original algorithm,which can guide the algorithm to ensure the correct search direction in the iterative process,and avoiding convergence to the local optimum due to the lack of global information in the early stage.Finally,using the Cauchy mutation to increase the diversity of the population and help the algorithm to jump out of the local optimum.Six test functions are set up to simulate the experiment and verify the optimization performance of the improved algorithm.The experimental results show that the DCFPA algorithm is effective in avoiding premature convergence and jumping out of local optimum.(2)An adaptive flower pollination algorithm(OTAFPA)based on opposition-based learning and t-distribution is proposed.Firstly,the opposition-based learning strategy is used to increase the diversity and quality of the initial population.In order to balance the global search and the local search,the dynamic switching probability is introduced on the basis of the original algorithm,which increases the flexibility and adaptability of the algorithm.Finally,the tdistribution variation is used to increase the diversity of the population,and help algorithm jump out of the local optimum.The simulation results of eight test functions are used to verify the global optimization ability,convergence speed and solution accuracy of the OTAFPA algorithm.(3)The improved flower pollination algorithm is used to solve the problem of user identification across social networks.By using the improved flower pollination algorithm to optimize the weights and thresholds of the BP neural network,the discriminant experiments performed on the processed experimental data sets verify the effectiveness of the improved pollination algorithm.