Utility Maximization In Wireless Multi-Media Network

In the past few years the wireless communication has witnessed a tremendous development, especially in multimedia traffic which help wireless networks users to obtain a great quality of experience. Providing Quality of Service (QoS) to various multimedia users according to their bandwidth requirement is an important resource management issue for wireless network. However, due to the limited resource and variable wireless link environment, the resource management in wireless networks is still a challenging problem. Bandwidth allocation is one of the most important subject in various kinds of resource allocation mechanism. Providing a network bandwidth allocation mechanism, which can be adjusted dynamically to cope with the various kinds of fluctuations in wireless network is of great importance. In this paper, we will focus on the bandwidth allocation mechanism in order to obtain the global maximization of the total users’ utility.So far, satisfactory results have only been developed along the resource allocation among non-real-time traffic, which is eased by the concave utility function of such traffic. However, adaptive real-time users have non-concave utility functions, which means traditional convex optimization methods are not suitable. Facing the above challenge, in this paper we describe the resource allocation problem into a NUM model, in which the objective function is formulated by the non-concave utility functions. Rather than to solve this non-convex optimization problem directly, we propose an approximate model for the original model, which is easy to solve by traditional optimization method. Alongside, we will analyze the relativity of the original model and the approximate model. we novelty create the competitive point for each user, which can reflect their utility increment and potential increase rate. Based on the optimal solution we obtained from the approximated model, we proposed a gradient type greedy iteration method to obtain the solution for the original problem. The performance of the proposed method is further evaluated via simulation results, which demonstrate that our algorithm can be applied into various kinds of adaptive real-time traffic.