Automated Negotiation in Agent Based Commercial Grids

Abstract

Negotiation is a joint decision making process among actors with competing goals. by automating negotiation processes using agent technology, the cost of this decision making process in terms of time and manpower can be effectively reduced. Competitive negotiation between decision makers is characterized by limited availability of relevant information. To succeed in negotiations, human negotiators use their intuitive and interpretational skills to leverage incomplete knowledge and identify mutually beneficial contract modifications. A major problem in the domain of automated decision making in competitive negotiations is that intuitive human decision making under incomplete information must be replaced by formal decision functions. The work presented here aims at providing formal strategies for multiple-issue negotiations that model knowledge-aware decision making processes, allowing an artificial agent to gain advantage over its counterpart. Apart from this purely competitive take, an additional negotiation protocol is presented that allows finding mutually beneficial contract modifications although preference information is withheld by the negotiating parties due to their competitive stance. The commercial exchange of distributed high-performance computational services (grid services) serves as an example area of application.