A Comparison Study for Different Number of Mesh Routers and Small Scale WMNs Considering Subway Distribution of Mesh Clients and Three Router Replacement Methods

In our previous work, we implementated a hybrid intelligent simulation system for Wireless Mesh Networks (WMNs) called WMN-PSOHCDGA by integrating Particle Swarm Optimization (PSO), Hill Climbing (HC) and Distributed Genetic Algorithm (DGA). We carried out many simimulation scenarios, but we used the same number of mesh routers in order to make a fair comparison. However, we found that in many cases some routers were concentrated in some areas. So, it is possible to use less number of mesh routers to cover all mesh clients. In this paper, we consider Subway distribution of mesh clients and samll scale WMNs. We carry out simulations for three router replacement methods: Linearly Decreasing Inertia Weight Method (LDIWM), Linearly Decreasing Vmax Method (LDVM) and Fast Convergence Rational Decrement of Vmax Method (FC-RDVM) and two scenarios where the number of mesh routers is 10 and 16. We evaluate the performance for these two scenarious considering three metrics: Size of Giant Component (SGC), Number of Covered Mesh Clients (NCMC) and Number of Covered Mesh Clients per Router (NCMCpR).