A primary challenge in wireless networks is to use available resources efficiently so that the Quality of Service (QoS) is satisfied while maximizing the throughput of the network. Among different resource allocation strategies, power and spectrum allocations have long been regarded as efficient tools to mitigate interference and improve the throughput of the network. Also, achieving a low transmission delay is an important QoS requirement in buffer-limited networks, particularly for users with real-time services. For these networks, too much delay results in dropping some packets. Therefore, the main challenge in networks with real-time services is to utilize an efficient power allocation scheme so that the delay is minimized while achieving a high throughput. This dissertation deals with these problems in distributed wireless networks.

A primary challenge in wireless networks is to use available resources efficiently so that the Quality of Service (QoS) is satisfied while maximizing the throughput of the network. Among different resource allocation strategies, power and spectrum allocations have long been regarded as efficient tools to mitigate interference and improve the throughput of the network. Also, achieving a low transmission delay is an important QoS requirement in buffer-limited networks, particularly for users with real-time services. For these networks, too much delay results in dropping some packets. Therefore, the main challenge in networks with real-time services is to utilize an efficient power allocation scheme so that the delay is minimized while achieving a high throughput. This dissertation deals with these problems in distributed wireless networks.

No / This paper proposes a novel call admission control (CAC) algorithm and develops a two-dimensional markov chain processes (MCP) analytical model to evaluate its performance for heterogeneous wireless network. Within the context of this paper, a hybrid UMTS-WLAN network is investigated. The designed threshold-based CAC algorithm is launched basing on the user’s classification and channel allocation policy. In this approach, channels are assigned dynamically in accordance with user class differentiation. The two-dimensional MCP mathematical analytic method reflects the system performance by appraising the dropping likelihood of handover traffics. The results show that the new CAC algorithm increases the admission probability of handover traffics, while guarantees the system quality of service (QoS) requirement.

This thesis treats of QoS (Quality of Service) in UMTS system. The term quality of service is frequently used, recently, however not only in computer networks, but also more and more in mobile networks also. Present system of 2nd generation (2G) GSM enables only limited exploitation of QoS function. System UMTS is a system of 3rd generation (3G) already and contains quality of service in greater measure which is ensured with number of algorithms and functions. Individual QoS functions enable effective utilization of radio interface, support maintenance of planned coverage and offer high spectral efficiency. Functions that cater to quality of service are called RRM functions (Radio Resource Management). There exist five basic functions in UMTS system that are Admission control (AC), Power control (PC), Load control (LC), Handover control (HC) and Packet scheduler (PS). These functions are responsible for control of network access, control of power, control of network load, control of handover and packet (bit rate) scheduling. Last three functions are usually collectively designated as Congestion control, control of network overload. UMTS layer model, differentiation of services, traffic classes and 3GPP concept are also discussed. In thesis the attention is piad to Admission control algorithm – control of network access – in the uplink direction and to different ways of load (in the cell) modeling. Individual approaches are simulated in MATLAB. A possibility of using algorithm in the real system is discussed in conclusion.