Automatic Service Composition Based on Enhanced Service Dependency Graph

Zhifeng Gu, Juanzi Li, Bin Xu

Abstract: Service dependency graph (SDG) is an AND/OR graph showing input output dependencies among service operations. As dependencies in an SDG are indirectly expressed by reasoning on data models used by service interface definitions, their re-usability and expressiveness are limited. In this paper, we propose an enhanced version of service dependency graph, namely SDG+. SDG+ enhances SDG with explicit dependency declaration, which expresses dependencies directly with static explicit declarations. Based on SDG+, we developed our automatic service composition algorithm for WS-Challenge 2007, which wins the championship of composition efficiency in the competition.