In most cases, QoS-driven global optimization is a non-linear 0-1 programming. Genetic Algorithms are superior in non-linear 0-1 programming and multi-objective optimization. However, encoding methods that many Genetic Algorithms (GAs) adopts are too complex or simple to apply to services selection. A novel Tree-coding Gas (TGA) is presented for QoS-driven service selection in services composition. Tree-coding schema carries the messages of static model of service workflow, which qualifies TGA to encode and decode chromosomes automatically, and keeps the medial results for fitness computing. Tree-coding can also support the services composition flow replanning at runtime effectively. The experiment results show that TGA run faster than one-dimensional Genetic Algorithms when the optimal result is the same, furthermore the algorithm with Tree-coding is effective for re-planning. © Springer-Verlag Berlin Heidelberg 2007.
CITATION STYLE
Gao, C., Cai, M., & Chen, H. (2007). QoS-driven global optimization of services selection supporting services flow re-planning. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4537 LNCS, pp. 516–521). https://doi.org/10.1007/978-3-540-72909-9_56
Mendeley helps you to discover research relevant for your work.