| The concept of radio spectrum reuse has emerged as one possible solution to the severely limited availability of new spectrum bands and the underutilization of the allocated bands. Therefore, in the future, it is expected to have users that belong to different access priority sets in a given wireless network, operating on the same spectrum band. In this dissertation, we study scheduling and resource allocation problems when users with different access priority to the shared spectrum resources interact either cooperatively or competitively.;First, we study optimal transmission scheduling of backlogged users under the spectrum leasing model. Here, we extend the opportunistic scheduling framework and propose a cooperative scheme based on decode-and-forward multihop in which secondary users may share the time slot with primary users. Cooperation is allowed only if either instantaneous utility or long term utilities of users is improved. In both cases, our proposed algorithms are shown to outperform non-cooperative scheduling algorithms, in terms of both individual and total expected utility. Using stochastic network optimization tools, we show that the performance of our algorithms is arbitrarily close to the optimal performance at the price of reduced convergence rate.;Next, we study user cooperation when users have bursty packet arrivals and varying connectivity. In addition, primary users have secrecy constraints with respect to an external eavesdropper. Secure stable throughput region is characterized for multiple cooperation protocols that incorporate packet relaying and collision generation. When channel state is known, we show that a collision generation policy can always provide gains over the non-cooperative policy while it is not the case for relaying policies. In addition, a policy that involves both cooperative relaying and collision generation may achieve inferior performance compared to an only collision generation policy, even if the cooperative relaying policy is better than non-cooperative policy. When channel states are unknown, we develop a cooperative collision generation policy that can outperform relaying policy under certain channel conditions.;Finally, we consider multiple access games with selfish users, where primary users' communication is confidential and secondary users may employ eavesdropping or jamming as a leverage to maximize their utility at the equilibrium. Users perform resource allocation to maximize their achievable rates minus transmission cost. Two types of decoders are considered. For the successive interference cancellation decoder, Nash equilibrium of the two-player game is shown to be Pareto inefficient. Then, it is shown that secondary users follow primary users where the Stackelberg equilibrium Pareto-dominates the Nash equilibrium, even if the eavesdropper channel is better than the main channel. In this scenario, we also study the unknown eavesdropper channel case and show that it is not always beneficial for the secondary user to hide the actual eavesdropper channel gain. For the single user decoder, the destination is set as leader in a Stackelberg game maximizing primary users' utility. Interestingly, the equilibrium reveals a recruiting process that turns a set of eavesdropping secondary users into helping jammers under certain conditions. |
