Application of Discrete Particle Swarm Optimaization for Service Selection in Grid

Abstract

Service selection plays a very important role in the grid middleware. Firstly, a service redundancy strategy is adapted and the corresponding target equation is designed for improving the QoS of the whole workflows in the grid. Secondly, considering of determining a optimal service schema virtually belongs to a combinatorial optimization problem, a resolving algorithm based on discrete particle swarm optimization to find a global optimum is proposed, and its encoding methods are given, In which, the suitable parameters in the algorithm are decided through the adjustment experiments. Finally, the results of some experiments show that the algorithm is feasible, and compared with some other algorithms, such as local optimization, greedy algorithm etc, the algorithm in the paper is more effective.