Novel Pre-pushing Scheme for Peer-Assisted Streaming Network based on Multi-leader Multi-follower Stackelberg Model

Peer-assisted streaming network plays an important role in people’s daily lives, for users can watch multimedia resources anywhere, anytime. In this kind of network, servers need pre-pushing certain resources to users for caching which can be further retrieved by other users, and then the network performance can be improved. A key design algorithm in pre-pushing scheme is that which user should be selected to be pushed and how to price to keep the effectiveness and fairness of this action. To address this issue, we turn to multi-leader multi-follower Stackelberg model which is a hierarchical game model dealing with the resource competing among leaders and followers. In our model, we set that users are leaders who own bandwidth and define the price of a unit amount of bandwidth. Servers are followers and each of them selects a best user for pushing according to leaders’ equilibrium results. We build utility functions for each server and user, respectively, and maximize them according to the built Stackelberg model. The numerical simulation demonstrates the effectiveness and fairness of our proposed pushing scheme. Further, the comparison between traditional pre-pushing scheme and our proposed one is simulated which also shows that our proposed one can make more payoffs.