Comparing and Evaluating Collection Strategies in Trust-Based Reputation Systems in Distributed Environments

Trust-based reputation system (TBRS) provides us an excellent decision-support tool for distributed environments. TBRS quantifies the trust relationships between entities based on the information that describes the entities´ past behavior, so it is significant for TBRS to guarantee the quantity, the quality and the reliability of the collected information. In distributed environments, due to the lack of a central authority that is responsible for collecting global information, it is still a challenging work to design an effective and efficient information collection (IC) strategy. A variety of TBRS have been proposed in previous papers during recent years, however, there are few works on comparing and evaluating different IC strategies in TBRS. In this paper, we propose a formal model that is mainly focused on IC strategies in TBRS, classify existing IC strategies into three categories and describe their main features respectively. We propose a simulation framework based on the formal model and implement it for comparing and evaluating the performance of different IC strategies. We analyze the results of simulation and present the strength and weakness of different IC strategies.