Research On Personal Decision Support Assistant Agent

The rapid development of information technology and Web has benefited people greatly, which also makes the decision environments even more complex. In order to be adapted to the ever-changing and complex environment, and to make decision efficiently and effectively, decision-makers are eager to be provided with personal decision support services. In such a context, how to satisfy the decision-makers' personal decision support requirements is becoming a hot problem. The personal assistant agents are playing a more and more important role according to its characteristics, such as intelligence, personality and cooperation.Personal Assistant ("PA" for short) is one of the central class agents. A PA is a software agent that acts semi-autonomously for and on behalf of the user, modeling the interests of the user and providing the required personal services to the user. In this dissertation, Personal Decision Support Assistant Agent ("PDSAA" for short) that belongs to PA is put forward. PDSAA helps the decision maker analyze decision problem, which provides information that is correlative with the decision problems and then recommends the personal solutions by participating in the user's decision process. PDSAA can alleviate the decision makers' working load effectively and thus can improve decision efficiency.In this dissertation, firstly, the reference model and the corresponding working procedure of PDSAA are presented. Then, as the core of the PDSAA the presentation and construction methods of user profiles are thoroughly studied. An ontology-based and weighted hierarchical presentation approach and a three-phase construction method for the user profiles are advanced. Based on the methods proposed above, how to provide the decision makers with personal decision support services by combining decision-making methods with user profiles is thoroughly discussed. The algorithm for recommending solutions is presented on the basis of user profiles and TOPSIS method. Finally, through using the most popular agent development platform JADE (Java Agent DEvelopment Framework), the author has designed and implemented a PDSAA prototype for supporting house-buying decision making to illustrate how the methods proposed in the dissertation work.