| In multimedia sharing networks such as YouTube, and Flicker, etc, users actively participate and interact with each other, which influences not only each individual but also the entire system performance. Successful deployments of multimedia sharing networks show that user cooperation helps provide efficient and highly scalable platforms for multimedia distribution. However, since users are selfish, their cooperation cannot be guaranteed. In this thesis, we aim to design incentive mechanisms to stimulate user cooperation and also optimize the system performance.;We then study an interactive multiview video streaming (IMVS) system, where an user can select one out of many available views for observation and switch views frequently. With the advances of multiview video coding techniques, users can cooperatively download videos even if they are watching different views. We then model user interaction as an indirect reciprocity game and formulate users’ decision making associated with their view switching as a Markov decision process.;In these two examples, our analysis shows that user behavior impacts the system performance significantly. Thus, we optimize our incentive mechanisms, which drive the games to desired stable equilibria, where users cooperate with each other and the system performance is maximized at the same time.;Without loss of generality, we use two multimedia applications as examples to show how to achieve our research goals. We first study a two-hop cooperative wireless multicast system, where after the base station broadcasts a packet, a relay node who receives the packet correctly helps forward it to the others. We model user interaction in this system as a multi-seller multi-buyer payment based game, where users pay to receive relay service and get paid if they help forward a packet. |
