关键词: 布谷鸟搜索, 莱维飞行, 旅行商问题, 群智能算法

Abstract: Cuckoo Search（CS） algorithm shows better performance in solving continuous problems, but the existing CS algorithm converges slowly and fails to reflect the characteristics of Levy flight. To solve the drawback, a new Genotype-Phenotype Cuckoo Search（GPCS） algorithm is proposed in this paper. In the GPCS algorithm, each city is encoded with random decimal called gene which the integer part is the city number, and the meanings contained in the genes are together decided by the decimal and the integer parts. The decimal part determines the access order of the city, the integer part represents the city number, and the two parts together constitute the neighborhood space of Levy flight, then selects the relocate or displace operation according to the flight results. The experimental results show that GPCS algorithm is not only better than other cuckoo search based algorithms, but is better than some state-of-the-art swarm intelligence algorithms too.

Key words: cuckoo search, Levy flight, traveling salesman problem, swarm intelligence algorithm