A SMALLER PRICE ON GLCA DYNAMIC SERVICE COMPOSITION ALGORITHM

The key factor affecting dynamic service composition is in the cost of services combined. The paper provides a model for service combination firstly based on dynamic service portfolio Candidate Service model, taking a directed acyclic graph. Moreover, according to the dynamic characteristics of service costs, the Greedy Local Correction Algorithm(GLCA) was used in the synthesis process. The algorithm has a high efficiency by reason of selecting the minimum number of current number every time, and the number of cycles is limited, the number of iterations related to depth of services. The comparison on time performance was made between GLCA algorithm and Dijkstra and Bellman-ford algorithm through simulation experiments, the result showed that the time performance of GLCA algorithm is better than other algorithms for the number of sub-services case increasing.