International Conference on Machine Learning and Cybernetics (ICMLC), 2010
Abstract:Optimization of service selection for composition is a critical but difficult issue in the area of service-oriented computing. This paper proposes an innovative approach to solve this issue. In the proposed approach, every abstract component service in the composition will be assigned a weight value to represent its importance through user assignment and execution path analysis, and then a weighted service flow will be generated. A service combination that includes selected concrete component services will be produced based on the weighted service flow and the Genetic Algorithm. The service combination will be of acceptable quality, and ensures the core component services that will most likely affect the whole service flow execution are more robust than other ones.