The composition of Web services is an interesting option for the creation of complex applications with a wide range of features. The inherent scalability of the Internet leads to a potentially large number of services that meet a particular feature. Moreover, these services are associated with different quality indicators, sometimes conflicting. Hence, it is evident the plurality of solutions for the same problem. Considering this fact, in this work, the composition of Web services is treated as a many objective optimization problem, and multiobjective evolutionary algorithms are used to solve it. An empirical evaluation is done and the results of different preference relation techniques are compared in different composite Web services scenarios. The results show the importance of applying many-objectives techniques to problems with more than three objectives.
Published in:
Neural Networks (SBRN), 2010 Eleventh Brazilian Symposium on
Date of Conference: 23-28 Oct. 2010