An enhanced QoS decomposition approach for efficient service composition

For each service class, there exist many alternative web services that provide the same functionality and different quality parameters, to find an optimal composition, most service selection approach in decentralized environment enumerate all possible combinations of service candidates. However, with the growing number of alternative services, those approaches will achieve a poor performance. To improve the performance, in this paper we proposed a QoS decomposition approach, that divide the global constraints into separate ones for each involved service class. With the separate constraints, each service class engages in service selection independently and simultaneously. Moreover, besides sequential pattern our approach also supports parallel pattern. Experimental results indicate that our approach performs efficiently and can achieve near-optimal solutions.