Service Selection Based On QoS

Web service has several characteristics, such as cross-platform, low coupling, as well as language-independent, and so on.it has become an important solution for e-commerce and distributed computing. Due to the complex business requirements, a single web service is not able to fully meet their needs, which thus make the service composition become necessary. There is always a large number of web services with the same functionality while different QoS(Quality of Service) on the internet, therefore, how to select web services for composition with the purpose of having the best QoS and also satisfying user restrictions of QoS has become an urgent problem to be solved.Existing work mainly focused on finding a combination of approximate solution, the reason is that this solution can approach to the optimal solution faster. One commonly used method is genetic algorithm, because it is simple, and is good for parallel search. However, with the expansion of web services composition scale, standard genetic algorithm selection process gradually showed its disadvantage, such as slow convergence, and the difficulty in converging to the global optimal solution.In light of the above issues, this paper propose to do following work.1) A fuzzy elite genetic algorithm(Referred FEGA) has been developed. It improves the genetic algorithm in terms of the convergence speed and the global convergence. By taking advantage of the impact of elite individual to the genetic algorithms, in FEGA, the population is divided into two sub-populations in each generation, one is evolved with the best individual of the population, while the other is evolved regularly. The division of the sub-population is accomplished by the fuzzy control based on the current operating parameters of the algorithm. Finally, the two sub-populations merged into the next generation of the population.2) Applying the FEGA algorithm to web services selection. A web service composition example Trip Planner verifies the validity of the application of FEGA algorithm to the web services selection. The results show that it can quickly and effectively selects the combination which meets the user QoS needs, and also solves the web service selection problem with global QoS constraints.