SIMULATION OF NON-COOPERATIVE POWER CONTROL GAME FOR WIRELESS COMA NETWORKS

Abstract

Power control is one of the most important aspects of radio resource management in CDMA systems. With the advent of multimedia services in wireless CDMA networks power control has become more important. An efficient power control scheme makes the users to use the radio resources more efficiently, thus increasing the capacity of the system, battery life of the mobile terminal and mitigating near-far effect. The power control problem should be modeled such that various services meet their respective Quality- of Service (QOS) requirements. QOS may be the required target signal-to-Interference ratio (SIR), acceptable bit error rate (BER) or utility, the level of satisfaction of the user, which is the function of SIR or BER. In this dissertation work, utility function is considered to be the measure for QOS requirement. The main focus in this dissertation is on data services. In the wireless CDMA network all the users like to have higher SIR value while transmitting at lower powers. So, they have to satisfy conflicting objectives. In the situations like this game theory is used to provide the solutions. So, power control problem is modeled as a game where every user is a player. In the power control game all the users act selfishly in a non-cooperative manner to improve their utilities. So, in this dissertation power control game is modeled as a non-cooperative power control game. The non-cooperative power control game leads to Nash equilibrium, which is generally Pareto inefficient. So, pricing is introduced to bring in the Pareto improvements in the Nash equilibrium. In this work simulations are done for the non-cooperative power control game with and without pricing and it is shown that there is Pareto improvement in the Nash equilibrium due to pricing. Simulations are carried out for various situations like single cell/multicell, fading/AWGN channel, noncoherent BFSK/DPSK modulation and heterogeneous/data traffic. From all the simulations it is shown that pricing leads to higher utilities and lower transmit power for all the users compared to no pricing. From simulations cost value is observed to be proportional to the bit rate. The behaviour of cost parameter is also seen by varying system parameters and it is observed that cost value increases with the decrease in cell size and with increase in spreading gain i.e., the maximum number of users in the cell. iv