A research of multi-objective service selection problem based on MOACS algorithm

As more and more cloud services are provided on the web, composite services will increase exponentially. How to select the optimal services for a given composite service is a important problem and is NP-Hard. This paper proposes a multi-objective service selection algorithm based on MOACS(Multi-Objective Ant Colony System) optimization algorithm for this problem. In this problem, two objectives are considered. They are the response time and cost QoS attributes of services. Every objective has its heuristic function and it is defined against the characteristics of this problem. Two levels pheromone updating policy: global and local, are defined in this paper. A combined state transition rule is designed. Finally, this algorithm is evaluated experimentally using different standard datasets, and made some comparisons with the MOACO algorithm in the problem. The results show that our approach is effective for solving multi-objective service selection problem.