Eigenvalue filtering for random opportunistic networks

For the large-scale wireless mobile networks, the topology or global state information directly affects congestion control, traffic control, quality of service and so on. Due to the dynamic change of the opportunistic network structure, the node can not perceive the current state of the network, therefore, it is important to improve the routing quality by designing a topology aware algorithm that can predict the network topology or global state parameters. According to the theoretical deduction and experiment verification of opportunistic network, some conclusions of the average degree and eigenvalue curve is given. Based on these conclusions, an algorithm with multiple measured values for eigenvalue prediction is proposed. The results of the error analysis indicate that the prediction accuracy is obviously improved compared with the traditional Kalman filtering algorithm.