An Agent Negotiation Model For Fuzzy Intangible Requirement

Going with Internet continually improved and electronic commerce comprehensively applied, how to negotiate rapidly and effectively is one of important research of intelligent agent. Intention of negotiation is that bilateral or multi-issue negotiators reach agreement and acquire interest from it. To reach agreement needs cooperation by negotiators, however, every negotiator wants to use less information to acquire more information that belongs to opponents, so they can adopt more active negotiation strategies to pursue maximum profit, they will select to hide respective private information. They only pride the fuzzy intangible information to negotiate each other. This situation will hamper the cooperation among negotiators. Under the real negotiation scenario, the buyer who is very smart does not give the clear and unambiguous requirements that must be ambiguous and intangible ones to the seller. The seller who is very foxy adopts countermeasure to only disclose parts of product's information to the buyer. It hides some product's information that can be considede as reward information associated with the product. The seller uses the rewards associated with products to make sale promotion. Especially, the seller never reveals the profit acquired from selling a product at certain price. Therefore, it is so difficult to deal with fuzzy intangible information that negotiation often leads to fail.Seller and buyer agents need copy with fuzzy intangible information. They must cooperate for making concession to come to an agreement. At the same time, they also make minimum concession and acquire more interest. Therefore, how to deal with buyer's fuzzy intangible requirements is the key to solve the contradiction.This thesis develops a negotiation model. Depending on the model, seller and buyer agents can adopt strategy associated with hiding information, offering, and making concession to deal with fuzzy intangible requirements. Negotiation under the model not only can come to an agreement, but also the outcome is Pareto optimal. In the negotiation model, the buyer agent uses the fuzzy intangible requirement notion to hide its private requirement information. The seller agent uses fuzzy tables and their associated distinctions matrixes to acquire some part of products' information as reward associated with these products and hides these reward information. The buyer and seller agents' offers are generated under decreasing payoff order and they only make minimum concession. Under finite negotiation process, the model will deal with fuzzy intangible requirements and come to a Pareto optimal agreement. The major research work in this thesis can be viewed as follow:1. The research base of this thesis is intangible requirement satisfaction problem (LRSP). Theconstraint satisfaction problem (CSP) is quoted to formalize LRSP. The constraint is used to formalize intangible requirement and constraint satisfaction is used to formalize intangible requirement satisfaction.2. At the base of LRSP, this thesis puts forward fuzzy intangible requirement satisfaction problem (FLRSP). In fuzzy mathematics, the cut-set technique is quoted to extend CSP to fuzzy constraint satisfaction problem (FCSP). FCSP is used to formalize FLRSP. The buyer can use FLRSP to hide its requirements' information.3. Adding priority into the intangible requirement that also include fuzzy intangible requirement can embody the buyer's preference. The buyer offers according to his preference.4. This thesis use fuzzy table and associated distinctions matrixes to find standard information of seller's products as rewards information according to comparing the distinctions of products. The seller will hide the rewards information associated with products and uses the rewards to promote products for saling.5. The acceptability extent formula that is adopted by the model provides a computablesolution for negotiation which lets satisfaction situation of the required product and saling situation that is influenced by product's retrictions and rewards from the unformalized and uncomputable form into the formalized and computable process.