Abstract: Current methods of service selection based on quality of service (QoS) usually focus on a single service request at a time, or let the users in a waiting queue wait for Web services when the same functional Web service has more than one requests, and then choose the Web service with the best QoSfor the current request according to its own needs. However, there are multiple service requests for thesame functional web service at a time in practice and we cannot choose the best service for users everytime because of the service’s load. This paper aims at solving the Web Services selection for concurrent requests and developing a global optimal selection method for multiple similar service requesters to optimize the system resources. It proposes the improved social cognitive (ISCO) algorithm which uses genetic algorithm for observational learning and uses deviating degree to evaluate the solution.Furthermore, to enhance the efficiency of ISCO, the elite strategy is used in ISCO algorithm. We evaluateperformance of the ISCO algorithm and the selection method through simulations. The simulation results demonstrate that the ISCO is valid for optimization problems with discrete data and more effective than ACO and GA.
Keywords: web service selection, ISCO, deviating degree.
Guiming Lu, Yan Hai, Yaoyao Sun