Research On Improvement And Application Of Lion Swarm Optimization Algorithm

With the rapid development of information technology represented by the Internet,the swarm intelligence algorithm as an important optimization technique is widely used in signal processing,production scheduling,mechanical design and many other fields,providing a practical method to solve the combinatorial optimization problems of complex engineering.Lion swarm optimization algorithm,as a new type of swarm intelligence algorithm,has the characteristics of fast convergence speed and strong optimization finding ability.The multi-modal nature of the algorithm’s behavioral mechanism allows it to better adapt to different situations and thus accomplish a more comprehensive search for optimization;and the diverse information exchange mechanism allows the excellent characteristics of individual lions to be learned by other individuals and thus be retained and propagated.However,the lion swarm optimization algorithm still suffers from the problems of slow convergence in the late stage of optimization search and the tendency to fall into local extremes.To this end,this paper focuses on the improvement mechanism of the lion swarm optimization algorithm and applies the improved algorithm to the parameter optimization of neural networks,with the following main research contents.1.To address the problems of the traditional lion swarm optimization algorithm with low diversity and easy to fall into local optimum,an improved lion swarm optimization algorithm based on Tent chaotic mapping and differential evolution mechanism is proposed.First,to overcome the drawback of uneven population distribution caused by random initialization of the lion swarm optimization algorithm,an adaptive factor is introduced to improve the Tent chaos mapping to initialize the population position,which improves the population diversity and uniform traversal while ensuring random distribution,and then improves the global search capability of the algorithm.Secondly,to address the problem that the algorithm is prone to local optimum and low convergence accuracy due to the position update mechanism limited to the internal lioness,we introduce the lioness position update perturbation factor combined with the differential evolution mechanism to realize the adaptive adjustment of the algorithm,enhance its ability to jump out of the local optimum,and improve the speed and accuracy of the search.The algorithm is tested on 10 benchmark functions of various types and compared with six classical swarm intelligence algorithms to verify the convergence speed,the accuracy and stability of the improved lion swarm optimization algorithm.Finally,the proposed algorithm is applied to optimize the initial weights and thresholds of the BP neural network and applied to the study of the house price prediction problem.The effectiveness and superiority of the improved lion swarm algorithm is further verified through extensive comparison experiments and analysis on two standard data sets.2.To address the problems of slow convergence and poor global search ability in the late iteration of the traditional lion swarm optimization algorithm,an adaptive lion swarm optimization algorithm is proposed that integrates the second-order parametric,information entropy and chaos search strategies.Firstly,to address the problems of insufficient traversal in the early stage due to the lion cubs blind selection strategy,the step size perturbation factor is too much influenced by the solution space and easily falls into local optimum,the dynamic step size perturbation factor is formed by fusing the second-order parametric and information entropy,and the adaptive parameters are used to dynamically adjust the selection probability of different behaviors of lion cubs by the number of iterations to effectively suppress the premature convergence of the algorithm.After that,a chaotic search strategy is used to adaptively adjust the search range on the basis of the initialized population of improved Tent mapping,and to improve the poorly adapted individuals through multiple neighborhood points of the local optimal solution to further enhance the algorithm’s speed and accuracy of the search.The convergence speed,the search accuracy and the ability to jump out of the local optimal solution of the improved algorithm are verified through the comparison experiments with three classical swarm intelligence algorithms on 18 benchmark functions.Finally,the effectiveness and good optimization performance of the proposed algorithm are further verified through extensive experimental comparisons on 2 standard data sets by optimizing the initial weights and thresholds of the BP neural network with the improved lion swarm algorithm and applying it to the classification problem.