Abstract: To enhance the search accuracy and global convergence rate of Cuckoo Search (CS) algorithm,a Hybrid Adaptive Cuckoo algorithm based on Local search strategy (LHACS) was proposed.In this algorithm,a local population was randomly generated around each current solution,the local optimal solution was obtained by sine cosine operator and the current solution was replaced by local optimal solution,which had improved the local search accuracy actually.The adaptive discovery probability and factor of search step were substituted for the fixed ones in CS to enhance the global convergence speed of the algorithm.The experiment of 25 classical high-dimensional benchmark functions showed that the LHACS was superior to CS in convergence rate and search accuracy,and it was applied to engineering optimization design and 0-1 knapsack problems,which verified the validity of the algorithm.

Key words: hybrid adaptive cuckoo search algorithm, local search strategy, sine cosine operator, global convergence speed, swarm intelligence algorithm

摘要： 为了提高布谷鸟算法的搜索精度和全局收敛速度,提出一种基于局部搜索策略的混合自适应布谷鸟算法。在该改进算法中,每个当前解的周围随机产生一个局部种群,利用正余弦算子的局部寻优能力得到局部最优解,并用局部最优解替换当前解,以提高局部搜索精度;同时采用自适应发现概率和搜索步长替代布谷鸟算法中的固定发现概率和搜索步长,以提高算法的全局收敛速度。对25个经典高维基准函数进行实验表明,所提算法在收敛速度和求解精度上优于布谷鸟算法,通过将其应用于拉压弹簧、三杆桁架设计和0-1背包问题,验证了算法的有效性。

关键词: 混合自适应布谷鸟算法, 局部搜索策略, 正余弦算子, 全局收敛速度, 群智能算法