IEEE Transactions on Mobile Computing
| Hi-index | 0.00 |
This dissertation takes a game theoretic approach to model non-cooperative relationship between the service providers and the customers in a competitive environment, and utilizes the game formulation to propose an novel framework to solve the resource management problem in code division multiple access (CDMA) wireless data networks, providing flexibility to support differentiated quality of service (QoS) to different classes of users at the same time. At the session (or macro) level, the admission control problem is formulated as a non-cooperative game between the service provider and the user(s) requesting admission into the network. Two modes of admission are considered: admitting the users one at a time, and in batch mode. The user behavior of migrating between service providers, which is called churning, is modeled with the help of utility functions for the players. It is shown that each instance of the admission control game has an equilibrium or a dominant strategy, which clearly defines the service provider's admission strategy. It is also shown that the batch admission mode yields better overall utility, at the cost of longer admission delay. At the medium access control (MAC) frame (or micro) level, a multi-rate CDMA system model is employed. In this model, different groups of users are allowed to have different subscribed transmission rates, and it is allowed to go below the subscribed rate when congestion happens at the air interface. The proposed optimal rate allocation strategy is guided by the user churn rate that is derived directly from the user's satisfaction (utility). The NP-Complete optimal rate allocation problem is solved by highly efficient heuristics that produce high quality sub-optimal solutions. The admission control and rate allocation schemes at two levels interact with each other through the use of the utility functions. In particular, the rate allocation decides on the instantaneous utility for the service provider and the user, while the admission control makes use of such instantaneous utility and ensures that the overall utility is maximized. The performance of the integrated admission and rate control framework is studied in detail with extensive simulations under realistic traffic models. Analytical and simulation results demonstrate that the proposed framework is able to improve the total revenue for the service provider as much as 45%. Moreover, the differentiated QoS is achieved for different classes of users, both at the session level and the frame level.