An Improved Evolutionary Multiobjective Service Composition Algorithm

摘要

Evolutionary multi-objective service composition optimizer (E3) is a recently proposed optimization framework for SLA-Aware service composition. It considers multiple SLAs simultaneously and produces a set of Pareto solutions. Two multi-objective genetic algorithms: E3-MOGA and Extreme-E3 provided by E3 have shown very good performance in comparison to NSGA-II. In this paper, an improved version of E3-MOGA, namely E3-IMOGA is proposed, which incorporates a fine-grained domination assignment value strategy. We evaluated our approach experimentally using dataset and compared with E3-MOGA and NSGA-II. It reveals promising results in terms of the quality of individuals and the time for finding all feasible individuals.