Modeling constraint-based negotiating agents

摘要

Most existing decision support systems (DSSs) are hard to fit satisfactorily into emerging working practices or organizational environments. Decision-making is becoming more pluralistic and less hierarchical, determined not so much by position in the corporate hierarchy but much by the argumentative and evidential value. Such decision-making can only be supported by those DSSs with negotiation support facilities. Intelligent negotiation agents can be used to model many decision-making negotiation tasks. Such negotiation agents are able to interact and negotiate with users and with each other. In addition, the newly emerging constraint agent technology provides a promising solution for such negotiation agents. In this paper, we present a model by applying constraint negotiation agents for DSSs. An object-oriented constraint language for modeling constraint agents is defined. Within our model, constraint DSS agents are able to reason cooperatively with users or with other agents. Negotiation among agents and users is modeled as a process of interactive constraint satisfaction. Efficient algorithms for the implementation of negotiation-oriented constraint satisfaction are also presented. As manpower costs become more and more expensive and as the negotiation workload increases, the future DSSs will play more important roles in the negotiation process. It is expected that our work will contribute to the future DSS design and development.