CL-DECCM-SA: A Cluster-based Delaunay Edge and Simulated Annealing Approach for Optimization of Mesh Routers Placement in WMNs

Abstract

Wireless Mesh Networks (WMNs) enable large and stable networks to be built at low cost through wireless communication among multiple mesh routers. There are many methods, algorithms and systems to optimize the placement of mesh routers. In our previous work, we proposed Coverage Construction Method (CCM), CCM-based Hill Climbing (HC), and CCM-based Simulated Annealing (SA) systems for the mesh router placement problem. We also proposed a hybrid system by combining Delaunay Edge and CCM-based SA (DECCM-SA) in order to increase the probability of finding distant mesh clients. In this paper, we propose a Clustering-DECCM-SA (CL-DECCM-SA) as an improvement of the DECCM-SA system. The proposed CL-DECCM-SA approach clusters the mesh clients and performs Delaunay triangulation on the average point of each cluster. After that the derived Delaunay edges are approximated to integer values and used as the range of possible mesh router locations for the CCM to generate the initial solution. We evaluated the proposed approach by simulations. For the simulations, we consider the evacuation areas in Okayama City, Japan, which is the target to be covered by mesh routers. The simulation results show that the proposed approach is able to place mesh routers near mesh clients that are clustered far away from each other and can cover all mesh clients.