In this thesis, we advance the fundamental theory of distributed joint codeword and power design and investigate how it can be applied in the context of dynamic systems. Mainly, this thesis proposes transmitter optimization algorithms for conventional CDMA models and cognitive OFDM systems to mitigate the interference presence. It takes into consideration the dynamic aspects of the network configuration where the user codeword and power have to be designed for systems with varying number of end-users and Quality of Services (QoS), while the system continues to transmit/receive information. First, an adaptive algorithm for interference avoidance improved with a power control mechanism is presented for distributed users in a CDMA wireless system. The algorithm is derived using a game theoretic approach in which separable cost functions with respect to codeword and power are defined, and joint codeword and power adaptation is formulated as a separable game with two corresponding subgames: the non-cooperative codeword adaptation game and the non-cooperative power control game. The algorithm can track variable target SINRs or variable number of active users in the system and is therefore useful for dynamic wireless systems with varying quality of service (QoS) requirements. Alternative approaches for decreasing the computational complexity are provided. Feedback channels that match to our incremental algorithms are presented. Second, we introduced and analyzed the uplink of SPA-OFDM system, where the unlicensed user can adapt its rate according to the upper layer applications. In order to improve the system rate flexibility and preserve the spectrum usage, the players will be the transmitted bits for which we define individual cost functions. The user optimizes different number of bits according to the upper layer applications. The transmitter optimizes the spectrum patterns while the powers are adapted subject to SINR constraints. We define an admissibility condition based on which the upper layers of the system will decide what applications can be run according to the maximum allowed transmitted power on each frequency band. It is shown that even if the optimization problem does not have a unique equilibrium point, the equilibrium will be optimal. For the optimal point, the sum capacity is maximized for the minimum power allocation. Finally, the idea of parallel optimization for distributed end-user is introduced as an additional mean to increase the convergence speed while decreasing the required amount of feedback information. Parallel iterations are implemented using gradient approaches for constrained optimization problems. Contrary to a general belief that gradient based algorithms decrease the algorithm convergence speed, the proposed method increases the convergence speed. Additionally, as the number of active users increases we observed more improvements. As a practical extension of my thesis, the GPS implementation proposes an adaptive algorithm which improves the robustness of indoor operation for weak signal conditions. Specifically, we design a blind MMSE receiver that can be implemented in GPS receivers as an additional loop that works during tracking mode to decrease the bit error probability. Moreover, reduced complexity adaptive algorithms are proposed.