Using Fuzzy Type-2 Inference Mechanism in Automated Negotiation Modeling

Intelligent agents are capable of negotiating with each other to increase their benefit in purchasing or any type of exchange process. They can offer some issues to each other to negotiate on. Negotiation can be either accepted or rejected by any of agents. In the negotiation models that have been presented so far, the issues and the evaluation of them haven considered as certain concepts and have been dealt with as crisp numbers. In this model, we claim that both issues and their evaluation are uncertain concepts, and we have used fuzzy type-2 membership function for representation and evaluation of the issues and fuzzy type-2 inference mechanism for decision making regarding the acceptance or rejection of the offer. In this work, the uncertainties of linguistic variables that are used to represent the quality or quantity of the issues have been modeled by type-2 fuzzy membership grades and the evaluation of them has been modeled by fuzzy type-2 inference engine. By this model, we can have a more human-like negotiating model, which is able to handle uncertain data and is more flexible. The fuzzy inference engine is a rule-based system, which can work, based on the rules that govern the behavior of the agents. The model has been implemented on the housing market bidding process between two agents i.e., seller, buyer, and the results have been demonstrated in this paper.