الفهرس 
COVEROnly 14 pages are availabe for public view
 |  | from | 162 |  |  |
| | | from | 162 | | |
Abstract
The current thesis is concerned about the implementation of an assessment scheme for the HWN and its performance evaluation. First of all, the thesis aims to establish the ability of having a complete analysis for different performance metrics in different RATs. The presented model has provided a service differentiation between new voice calls and handoff voice calls as well as new data sessions and handoff data sessions in the same RAT. This is done via the deployment of cooperative CRRM technique.
The presented work uses the multi-threshold bandwidth reservation scheme for multi services in both of homogeneous and heterogeneous wireless networks, which acts as an extension for the GC scheme. Performance measures are determined by modelling the system via a birth death process and solving recursively the linear equations derived from the Markov chain.
In certain cases existing calls can be terminated to admit offered handoff calls. The user’s priority of a forced termination is classified by its call holding time and call residence time. The system performance is investigated in terms of handoff voice forced termination probability, new voice call blocking probability, data call blocking probability and bandwidth utilization. The current thesis has two main contributions.
The first one is the determination of the optimum cost using the weighting factors to evaluate the system performance under multi-threshold reservation scheme.
It’s found that for having minimum cost, which corresponds to certain weighting factors, this leads to minimize the blocking probabilities and so improve the system performance.
The results can be summarized as follows:
• Differently from the previously published work, the current thesis has presented an analytical model which not only has the ability to be more tractable, accurate, and stable but also it has other features such as the possibility of finding out the performance for different cost models.
• A comparison between dual and multi-threshold schemes has shown that the system performance has been improved by the use of multi-threshold scheme.
• The guard channels for handoff calls can decrease the handoff blocking probability at the expense of new call blocking probabilities. However, the amount of guard channels to be reserved should be carefully selected. Excessive channel reservations may incur an unacceptably high level of new call blockings. This problem can be alleviated by adjusting the weighting factors , and , that determines the reservation threshold N. These optimum values had been deduced throughout the course of the presented work.• The obtained results show that there will be an increase in the blocking probability as the cell radius increases. This phenomenon arises due to the increase of the collected voice traffic as a result of enlarging cell coverage area. This leads to an increase of the blocking probabilities.
The other contribution in the thesis is to address the bandwidth allocation issue for an integrated voice and data in HWN. This has been done in conjunction with the QoS guarantee. The presented model has the novelty of combining both the multi-threshold priority scheme and the HWN with multi-dimensional Markov model. from the results, the following conclusions can be summarized:
• The proposed analytical model has solved multi-dimensional Markov model using a multi-threshold reservation scheme with a complete analysis for different performance metrics in different RATs. This has been done for service differentiation between new voice calls and handoff voice calls as well as new data sessions in HWNs.
• The presented model may be used as a generic model to investigate different traffic types of HWN. The obtained results have shown a great improvement in the performance metrics. This has been due to the deployment of multi-threshold algorithm.
• Multi-threshold bandwidth reservation scheme has shown large difference between its obtained KPI’s in GC scheme. This is due to the segmentation procedures that had been deployed to differentiate between the new and handoff traffic. This differention is reflected on the obtained RRM policy and improve the overall radio resource utilization.