الفهرس 
1. INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 176 |  |  |
| | | from | 176 | | |
Abstract
Low cost wireless routers are revolutionizing the way that people connect to the Internet. The ease of deployment and the freedom in the ability to connect have made these wireless routers ubiquitous. Wireless mesh networks extend the connectivity area of mobile devices beyond
the limited range of a single access point. These networks can be easily deployed inside a building, campus, large geographical area, or disaster site without requiring every access point to be physically connected to the Internet. They are also affordable when implemented with off-the-shelf low cost wireless routers. Wireless mesh networks consist of mesh clients, mesh
routers and gateways. Mesh routers have minimal mobility and form the backbone of wireless mesh networks. Gateways connect to the Internet using wireline links that supply Internet access to users. Due to the limited wireless channel bit rate, multiple gateways are usually required, which requires a budget and takes time to set the network up. In this thesis, we study
the network topology design and the gateway arrangement for minimizing the construction cost of a wireless mesh.
The interaction between computer science and optimization has yielded new practical solvers for global optimization problems, called metaheuristics. The structures of
metaheuristics are mainly based on simulations of nature and artificial intelligence tools. Metaheuristics mainly invoke exploration and exploitation search procedures to diversify the search all over the search space and intensify the search in some promising areas. Therefore,
metaheuristics cannot easily be entrapped in local minima. However, metaheuristics are computationally costly due to their slow convergence. One of the main reasons for their slow
convergence is that they may fail to detect promising search directions especially in the vicinity of local minima due to their random constructions. In this thesis, we will use a very diverse family of metaheuristics, namely simulated annealing, tabu search, genetic algorithms, ant colony optimization, and a hybrid algorithm that combines genetic algorithm with simulated annealing, to solve the Wireless Mesh Network (WMN) design problem
considering some quality of service (QoS) constraints.
In this study, we have evaluated the efficiency of the proposed algorithms; many experiments are conducted and compared with the well known algorithm (Predefined Gateway Set Algorithm). Each algorithm is distinguished in a specific aspect, depending on
its characteristics. The obtained results illustrated that proposed algorithms are more efficient.
For example, tabu search has the best network cost, genetic algorithm has a balance between minimizing the network cost and other constraints, simulated annealing and Predefined Gateway Set Algorithm have best running time, ant colony optimization has less number of
gateways in low capacity network, and hybrid algorithm has good minimum values of network cost especially in large size network. We have also evaluated the effectiveness of
maximum tolerable delay, link capacity, and maximum allowed antennas on minimizing network cost. Finally, we have searched around the optimal values of the relevant parameters in each algorithm.