| P2P (Peer-to-Peer) network technology has been widely applied in the network filesharing, distributed computing, e-commerce, etc., which has become a hot research topic inthe field of computer networks. However, due to the openness, anonymous, randomness,and network nodes dynamic characteristics of the P2P network, it also brings a series ofproblems such as covert communication, privacy protection, node collusion and fraud, etc.It is of great significance for the further development of P2P technology to introduce theconfidence-building mechanism in the P2P network, to built the trust model, to establish atrust relationship between peer nodes, and to build a trusted network interactionsenvironment. This thesis studies and analyzes the most representative P2P network trustmodels at home and abroad systematically, elaborates the characteristics of these modelsand points out their ascendancies and weaknesses. The current studies of P2P network trustmodel ignore the network dynamics and the calculation trust is not accurate. Because ofant colony algorithm having the characteristics of adapting to the dynamic networkenvironment and strong optimization abilities, this thesis proposes a new P2P trust model:trust model for P2P networks based on the improved ant colony algorithm (TMACA).The main work of this thesis as follows:(1) This thesis summarizes and analyzes the most representative eight P2P network trustmodels, and points out the advantages and weaknesses of existing models.(2) This thesis proposes an improved ant colony algorithm, which improved in threepoints: changing the methods of the local and global pheromone updating, redefining themeasure for the quality of the path that the ants found. Ants looking for the most trustedserver node in the network, they are guided by the pheromone traces which left by otherants going through the path. Therefore, the more access of the ants go through the path, themore obvious pheromone will left on the path, which will attract more ants to choose tofollow this path. Finally, the algorithm will converge to the best path that the most antshave gone through.(3) Applying the evolutionary bionic algorithm based on ant colony system to the P2Pnetwork system, this thesis proposes a new P2P trust model: trust model for P2P networksbased on the improved ant colony algorithm (TMACA). In TMACA, the traces ofpheromone left by the ants correspond to the amount of trust that a node has on its neighbors. The main reason of choosing to use the ant colony system to solve trust valuefor the P2P network is that, the high adaptability of the algorithm in a dynamic networkenvironment. Through this mode, the nodes in the P2P network can interact in a higherlevel of trust, because the pheromone remaining on the network tells us the reliability ofeach node in P2P network.The main innovations of this thesis as follows:(1) The improved ant colony algorithm has improved the original basic ant colonyalgorithm, and made improvements in the way of the local and global pheromone updating.This algorithm greatly improves the ability of finding optimal solutions and the globalconvergence.(2) The improved ant colony algorithm is applied to P2P network system, and constructsa P2P network trust model based on the improved ant colony algorithm. TMACA canenable nodes to select the best server to interact with, in terms of being the mosttrustworthy, at a very high probability in a dynamic network environment. This model cansolve the problem that the existing P2P network trust models ignore the network dynamicsand trust calculation is not accurate.The simulation experiments for TMACA haves been designed in three kinds of networkenvironment: static network, dynamic network and oscillating network. The results of theseexperiments show that TMACA can effectively improve the success rate for the node toselect the best server for interacting in the P2P network, and TMACA has a higher abilityto adapt in the dynamic changes in the P2P network environment. |
PDF Full Text Request