Research on transaction web service selection algorithm in WSC

Web service composition (WSC) needs transactional support to guarantee its consistent and reliable execution. To improve the validity of service selection, a transaction service composition model (TSM) is proposed. The model associates the quality of service (QoS) with dependencies between transactional states and time. Based on the model, a parallel algorithmlltransaction QoS selection algorithm (TQSA) is developed, which combines genetic algorithm (GA) and service compensation computing, and can be applied to dynamic service composition due to its mechanism of selection and evaluation performing at the same time. To improve the success of composition, an algorithm I-TQSA is further developed by optimizing evolution strategy. Experiment results show that TQSA resolves the QoS-aware selection NP-hard complete problem, and I-TQSA further improves the effectiveness of selection and reduces composition time.