A multi-attribute utility theoretic negotiation architecture for electronic commerce
AGENTS '00 Proceedings of the fourth international conference on Autonomous agents
| Hi-index | 0.00 |
A refined model for automated negotiation with reference to how we human perceive different concepts for the same product is examined in this paper. In multi-agent e-commerce environments, the requesting proposal from an agent is usually composed of attributes and weights that define the criteria and the preferences. Such a proposal is then posted to an electronic market place for matchmaking. When an agent finds no exact match, the proposal will be taken back, revised and re-submitted again. This process will continue as long as it takes. In most scenarios, we suppose that the requesters would already have the alternatives in their mind in case of no exact match. The concept-based model that we proposed allows requesters first specify their requirements and the alternatives altogether, in one proposal, before it is sent out to the e-marketplace. The alternatives are defined according to the different concepts that can possibly be perceived on a product. Different attributes may lead to different concepts under which the values of most attributes and weights may have to be changed. It is believed that by incorporating with the alternatives in the proposals, negotiation time could be shortened by saving some communication costs.