A distributed agent-based reputation framework enhancing trust in e-commerce
ASC '07 Proceedings of The Eleventh IASTED International Conference on Artificial Intelligence and Soft Computing
On desideratum for B2C E-commerce reputation systems
Journal of Computer Science and Technology - Special section on trust and reputation management in future computing systmes and applications
| Hi-index | 0.00 |
Reputation management schemes have recently emerged as a mechanism for improving trust and security in peer-topeer networks. A new entrant in a network with reputation management can either be given the benefit of doubt and be treated like a trusted (or semi-trusted) peer until it misbehaves or it can be assumed to be untrusted and have to earn the trust of others. The former case provides an incentive for misbehaving peers to cast off their old identity and assume a new one regularly. The latter case presents the problem of new peers being unable to bootstrap and to gain a foothold in the network. In this paper, we present a mechanism in which existing peers can choose to "lend" part of their reputation to new peers that they know in order to give them a start. This mechanism mirrors the real world where a reference or an introduction often gives one a foothold into an otherwise closed group. If the new peer behaves well, the old peer is rewarded for adding value to the group and if the new peer misbehaves the old peer loses the reputation it lent. In this way, existing peers have a stake in introducing new peers into the network and get a return on their investment for introducing honest peers.