Abstract: With the rapid growth of Web services in cloud computing environment, users make new demands on dynamic service composition, that quickly selects composite service which can meet multi-constraints. According to the demand, a multi-constraint service selection method based on decomposition of global QoS was proposed. A new cultural genetic algorithm was designed to decompose the global QoS constraints, and services were filtered by combining with the functional constraints between candidate services. An adaptive replacement strategy was used to dynamically adjust QoS constraint boundaries, so as to ensure the most appropriate composite service obtained by local selection. Experimental results show that the method can obtain nearly optimal composite service with a low time cost, which meets the dynamical and real-time features of service selection.

Key words: decomposition of global QoS, cultural genetic algorithm, functional constraints, adaptive replacement, composite service