A Reliable Fuzzy Theory based Reputation System in Grid

Abstract

Trust is a fundamental concern in Grid environment. Behavior trust that varies with time is indispensable in the trust system. Because of the fuzzy nature of behavior trust,  it is more appropriate to adopt fuzzy logic to express and compute trusts and reputations than adopt probabilities approach. A new reliable fuzzy theory based reputation system in Grid environment is proposed. By variable weighted fuzzy comprehensive evaluation, Direct Trust can be gotten and the affection of the ill factor have low value can be highlighted by give it a larger weight and then Grid entities can avoid choosing service providers with some defective. By derivation and combination of trust, Reputation can be obtained. Expert’s experience is used to set and simplify fuzzy rules. A distributed hash table based Dual Two-Layer Chord Protocol is proposed to store and retrieve reputations.  Simulation results show that entities in Grid can use resources or deploy services more securely in the support of  the reliable fuzzy theory based reputation system.