Research On Recommendation Methods Of Pareto Based On QoS Of Web Services Composition

In recent years Web services has been in a stage of rapid development, the number of Web services which can provide the same or similar functional properties is also increasing, so the number of research on the Web service composition QoS is also increasing. Web service composition based on QoS is a typical NP-hard problem, when the number of candidate services is large, how to choose a combination of services has become particularly complex.In this paper, to solve the shortcomings of the existing Web service composition QoS optimization problems, I do the work mainly from the following aspects.First, do the analysis of the previous Web service composition for solving optimization QoS problems. Most methods in dealing with such issues tend to follow a certain individual QoS inertia weight function or transform multi-objective optimization problem into a single objective optimization problem, these methods solve the multiple targets problems which can not be coordination to a certain extent, but the results can not objectively reflect the real needs of users. So in order to solve Web service composition QoS Optimization problem, the paper considers regarding the multiple objectives as a target vector simultaneously and considers the magnitude of each goal as specific evaluation results into account. Because of the essential characteristics of multi-objective optimization problem, multiple targets is often difficult to achieve optimal simultaneously with each other, so multi-objective optimization problem should no longer treated as a single objective optimization problem to simply get the best key. Therefore the paper introduces the concept of Pareto set, combines the multi-objective optimization solution process with the Pareto dominance, making the results obtained has strong Pareto dominance, and regards these highly dominant solutions as the optimization results.Second, carry out improvements on the standard particle swarm algorithm on the basis of the analysis of the specific Web service composition QoS optimization problem. As the standard particle swarm algorithm is mainly applied to solving continuous space optimization problem, and the Web service composition QoS optimization problem is a typical discrete space solving problems, so improve the standard particle swarm algorithm to discrete particle swarm algorithm, and change the original particle flying into the corresponding beating way. For PSO’s easy falling into local convergence properties, introduce the genetic algorithm mutation strategy and the ant colony algorithm’s pheromone thoughts into the new formula, so that the new PSO algorithm combines the advantages of the two algorithms.Finally, give a typical Web service composition model sequential structure model as an example, use the way of newly proposed particle swarm algorithm with Pareto recommendation to solve the Web service composition QoS multi-objective optimization problem through simulation experiments. By the relevant analysis of the experimental results, get the main influencing factors of Web service composition QoS optimization problem solving speed and by comparison with the existing methods to verify the feasibility and effectiveness of this method, highlighting the proposed method solving process efficiency advantage in certain circumstances. So provides a feasible way to solve the Web service composition problem QoS optimization.