Minglun Ren

Title: Service Composition Model Based on Improved GSA Algorithm: A Social Network Relationship Analysis Approach

Abstract: Internet of Services is an umbrella term used to describe several interacting phenomena that are shaping the future of how services are provided using the Internet. Services are characterized by intelligence, socialization, and personalization. They inter-relate, interact and cooperate with each other to realize certain goals. Social relationship plays an important role when services collaborate with each other, a role which has not been adequately investigated in previous research. The existing service composition methods consider functional qualifications and Quality of Service (QoS) as major consideration in service selection. It is difficult to adopt them to situations where interactive collaboration is required and social relationships between service providers play an important role in ensuring effective resources, information and knowledge hand-off in the process. In this paper, a service composition method based on weighted synergy network is proposed. Based on service interaction data, we construct the service social network which represent five kinds of relationships, namely Interactive transaction, Co-community, Physical distance, Resource-related and Social similarity. Based on these relationships the service synergy is derived. A service selection model that maximizes the overall synergy effect based on collaboration requirement is presented. An improved GSA algorithm that uses two-way learning, population update and group interaction based speed update mechanism is developed to solve the model. The model and improved GSA algorithm performs better and is validated through simulation experiment of intelligent automobile cloud manufacturing. The optimal service scheme (composition) in line with the actual manufacturing operation situation is obtained.