Kwok University of Hong Kong Trusted P 2 P Transactions with Fuzzy Reputation Aggregation

The Internet has enabled e-commerce and e-auctions for online commodity exchanges among strangers worldwide. Many companies now offer eauction platforms to facilitate such transactions, including eBay, Amazon, uBid, and Yahoo. A growing number of online commercial transactions occur in a peerto-peer (P2P) environment, which does not require a central authority to mediate such exchanges. Rather, participating peers can sign in and out asynchronously at will and perform their transactions point-to-point or point-to-multipoint anonymously. Thus, a business or enterprise must have an effective reputation system to help users locate trustworthy partners and exchange commodities securely with confidence. A reputation system’s effectiveness depends on the trust model adapted. Most existing trust models are developed for general-purpose P2P applications that emphasize distributed file-sharing1 (see the “Related Work in P2P Reputation Systems” sidebar). Little work has been done in establishing trust in e-commerce applications. Most notably, Daniel Manchala has assessed some trust metrics to quantify the trust among e-transaction intermediaries.2 Li Xiong and Ling Liu proposed the PeerTrust system for trust management.3 At the University of Southern California (USC), we aimed to develop an effective and efficient reputation system based on a fuzzy-logic approach, leveraging fuzzy-logic’s ability to handle uncertainty, fuzziness, and incomplete information