Multi-objective service compositon optimization using differential evolution

In order to handle the service QoS optimization problem, how to find out an effective scheme is the aim of our study. In the area of service-oriented computing, above problem is a typical NP-hard: assuming that given a business process, including a set of abstract service and a series of specific services, each series of specific service implementation of the corresponding each abstract service, how to find out the optimal combination of the specific service is our goal. For complex workflow, most recent research shows that the genetic algorithm is the best way up to now. Based on differential evolution (DE), we raise a new method. The experiment showed that the algorithm´s convergence speed is faster than the existing genetic algorithm.