Game-theoretic modeling for resource allocation in relay-based wireless networks

Game theory, originally invented to investigate the complex behavior in economics, has lent itself to a plethora of areas such as politics, biology, sociology and computer science nowadays. With its effectiveness in modeling the complicated behavior among different players, it is envisioned that it will play a significant role in the field of wireless communications and networking. In this dissertation, the author focuses on investigating some challenging resource allocation problems in relay-based wireless networks, a special type of wireless networks in presence of relays. To properly model the mathematical relationship between different network components and achieve a stable state where all components are satisfied with the outcome, different game-theoretic models are tailored to cope with different network scenarios. The specific issues the author addresses in this dissertation are summed up as follows. Firstly, the problem of price competition between different service providers (as owners of the relays) in wireless relay networks in presence of mobile users is considered. The payoff of a service provider is defined to be a function of both the relay price and the frequency that the relays are selected. To tackle the problem, the author designs a hierarchical framework via applying a Stackelberg-game model, where dynamic relay selection for mobile users and price competition for service providers are jointly taken into consideration. At the lower level, the author investigates the relay selection problem for the mobile users under given relay prices. It is formulated as a Markov decision process (MDP) problem with the objective to minimize the mobile user’s long-term average cost, and to be solved by applying the linear programming (LP) technique. At the upper level, the author studies the game of setting relay prices for the service providers, with the knowledgement that the mobile users will make relay selections based on their given prices. Nash equilibrium is obtained as the solution. The obtained results can help to provide a guidance for service providers to compete for providing relay services. Secondly, the author investigates the problem of power allocation for secondary users (SUs) in cognitive relay networks, where the SUs are involved as cooperative relays in a primary user’s (PU) communication. Opportunistic channel access scheme is considered, which motivates the SUs to use some relay power to speed up the PU’s transmissions. Consequently, the PU’s buffer will be depleted faster, resulting in more transmission opportunities for the SUs. Nonetheless, due to the energy limitation, less power can be used for an SU’s own transmissions if too much power is consumed on relaying. To address this tradeoff, the author designs power allocation strategies for both single-SU and multiple-SU cases. The former is formulated as a utility maximization problem, while the latter is modeled as a non-cooperative game. The existence and uniqueness of the Nash equilibrium (NE) are proved, indicating that a stable network state can be achieved where no SU can benefit more by unilaterally changing its power allocation strategy when the other SUs keep their strategies unaltered. Thirdly, the author studies the problem of cooperation formation for downlink data transmission in mobile infostation networks, where content distribution is achieved by exploiting the opportunistic contact among mobile users and infostations. Focusing on such a network scenario, the author discovers that the cooperation among infostations and mobile users can improve the network performance in term of the expected delay of content distribution. More interestingly, jointly taking both cooperation among infostations and that among mobile users into account may achieve even better improvement in performance, which inspires the author to propose a hierarchical cooperation formation model to analyze the bi-level cooperation. Coalitional formation game model and network formation game model are properly applied to obtain the stable structure of the bi-level cooperation, and an implementable distributed algorithm is proposed. Through extensive numerical experiments, the proposed algorithm is shown to be highly effective in obtaining the stable hierarchical cooperation structure.Finally, the problem of content delivery in relay-based publish-subscribe (pub-sub) networks is discussed. For the sake of optimizing the content delivery process from the content provider (CP) to the subscribers via relays in between, the CP plays a vital role in the strategy design. Modeling the pub-sub network using a tree-structured topology, tandem queueing model and absorbing Markov chain model can be applied to derive the quality-of-service (QoS) received by the subscribers (in term of the delivery ratio within deadline). It is observed that the content delivery problem has two major characteristics: (1) the CP always has the capability to control the content delivery process and is pursuing the maximum profit by strategically designing the “rights” and “obligation” items for the subscribers; (2) the subscribers are also self-interested so that they will attempt to achieve the highest benefits of themselves, however, their actions are constrained by the possible choices provided by the CP. Based on the above problem characteristics, contract theory is adopted to reach an economically optimal solution. The numerical results verify the effectiveness of the contract-theoretic approach in maximizing the CP’s profit, and the capability to ensure the satisfaction of the heterogeneous subscribers.DOCTOR OF PHILOSOPHY (SCE