A goal-driven and content-oriented planning system for knowledge-intensive service composition

With the rapid development of Cloud computing, social computing, and Big Data, the service composition in support of increasing knowledge-intensive innovation activities across enterprises and organizations sets a new challenge of dealing with diversified users' needs and changeable structures in process. Automatic service composition mainly makes use of artificial intelligence planning techniques to address these challenges. But current research focuses on separate aspects of these challenges, and special effort should be made to put goals, domain knowledge and activities into a unified perspective. This paper proposes a goal-driven and content-oriented automatic service composition method based on Hierarchical Task Network. First a unified planning domain knowledge modeling approach is given regarding goals, contents, tasks and their relationships, then a planning framework and corresponding algorithms are designed by extending Simple Hierarchical Ordered Planner 2, finally a loosely-coupled, extensible and flexible planning system is implemented and the experiment proves its feasibility and features.