Propagation And Aggregation For Trust And Reputation Based On Max-aggregation

The calculation of trust and reputation plays an important role in the current network. After a certain development, the technology has produced a series of theoretical results. The algorithm of trust propagation and aggregation is one of the classic algorithms for computing trust and reputation, combining with the principle of the interaction among people; it has been widely used in computing reputation. This paper discusses the direct reputation calculation method, and proposes a new trust propagation and aggregation algorithm, Max-agg algorithm, there the main work are as follows:First of all, in further study of the related theory about trust, reputation and P2P network, we find that the consideration about direct reputation value is not comprehensive, the accuracy is not enough. In order to overcome the existing defects, improve accuracy, this paper proposes a new trust reputation calculation method by introducing multi-dimensional and multi-attribute thoughts, that to ensure the calculation accuracy and comprehensiveness.Secondly, We put forward the optimal path, in this case, the optimal path is the path that can get closest true value after transmission, it can improve the transmission efficiency, when there are many path to get the witness node, we need to select a optimal path as the propagation path in the spread of trust, the experiments can prove that the optimal path for trust propagation can improve the accuracy, and have higher efficiency..Thirdly, We propose the Max-aggregation algorithm, the algorithm is mainly for the nodes that do not have directly interact, and looking for a credible path to get the indirect reputation value, then will get indirect reputation value to aggregate the forecast value, in the end we decide whether to interact with this node according to forecast value. Through the study of trust propagation and aggregation between the nodes, the final predicted value is obtained by the optimal path, that what we need.Finally, we prove the study though the detailed experiments in this paper, through comparing with the existing algorithm, we analyzed he result of the experiment in detail, which proves the correctness the execution efficiency of the algorithm...