Character processing for intelligent agents in the electronic market using fuzzy logic

This paper presents the design of fuzzy electronic markets based on several agents. The character of the buyer and seller agents influences their behavior in the market. Various factors play an important role in the precise, real, flexible and attractive design of a market. This research aims to model the market and the character of intelligent agents based on fuzzy logic. In the market, sellers are known with different titles or credits. Results confirmed that sellers with high levels of personality would gain more credit than other sellers, consequently, they earn more sales and profit. In this model, when a seller does not tell the truth about his product, the customer might also suspect the honesty of the seller about the quality of other products. In fact, the seller is recognized as a non-reputable person who possibly conceals the truth regarding the quality of his products. Therefore, the costumer will focus on sellers with high reputation for the future purchases. If the costumer finds no reputable seller, he will buy from those sellers whose credibility has not yet been evaluated (disreputable seller). He only purchases again from the non-reputable sellers if he could find neither the reputable seller nor the disreputable seller. Salesmen can offer promotions for attracting customers which is the most important goal of sellers. One of the promotions is giving discounts to buyers who have made more purchases from specific sellers or have brought more profit for the seller. In the proposed model, the discount was an important factor in attracting customers, which was achieved by sellers with high personalities. This feature was not included in previous models, and we have implemented this model with Aglet, and MATLAB. Results indicated that fuzzy agents modelling buy/sell based on their personality are more satisfied than sell/buy agents using only fixed bids.