为了提高标准布谷鸟算法的种群多样性和全局搜索效率,将多策略差分变异过程引入布谷鸟算法中.在布谷鸟的宿主发现过程中借助多策略差分操作来提高种群的多样性,同时,改进的布谷鸟在算法新解选择中增加排队优选机制,与贪心算法相结合以减少局部极值的不良吸引,加快搜索进程.将改进的布谷鸟算法应用到粒子滤波中,用布谷鸟的鸟巢来表征粒子,通过模拟布谷鸟群体搜索巢穴位置的过程来优化粒子分布.实验表明,改进的智能优化粒子滤波算法有效提高了粒子多样性和非线性系统状态的预测精度,并能在粒子数减少的情况下保持稳定估计.%Cuckoo search algorithm (CS) is a valid bio-heuristic algorithm,which has been extensively applied to solve the optimal problem in actual engineering projects,due to the advantages of simplicity,few parameters and easy implementation.In order to improve the population diversity and global search efficiency of the standard CS algorithm,the different mutation processes of an improved difference evolution algorithm was introduced into the cuckoo algorithm.In the different mutation processes,the multi-strategy associated with random walks method of the CS algorithm was used to optimize the host discovery process.With the multi-strategy difference mutation operation,the diversity of the cuckoo population was improved in the process of the cuckoo searching.Meanwhile,in the improved cuckoo searching,the queue optimization mechanism was added to the new solution selection,combining with the greedy algorithm to reduce the attraction problem of the undesirable solution and speed up the search process.In addition,the improved cuckoo algorithm with multi-strategy different mutation processes was applied to particle filtering.The particles were characterized with the cuckoo nests,by simulating the process that the cuckoo groups searched the nests to optimize the particle distribution.The experiment result showed that the improved particle filter can improve the prediction accuracy of particle diversity and nonlinear system state,and it can keep a good robustness and stability in the case of the particle number decrease.
展开▼
