Research On The Rechargeable Node Placement Problem In Green Wireless Mesh Networks

With the rapid development of wireless communication networks and the huge explosion of the Internet applications, Wireless Mesh Networks(WMNs) have become a popular access solution suitable for the “last ten miles” and have been widely developed in realistic scenarios. However, operating a large scale of WMN brings about high expenses on energy consumption due to the use of wired electricity as for node power supply. To solve the challenges of ever-increasing energy consumption, rechargeable mesh routers that can harvest their energies via solar or wind power have attracted a lot of attention. In this thesis, we study the node placement problem for a green WMN consisting of rechargeable routers, such that the total deployment and operational costs can be minimized, while guaranteeing user Quality of Service(QoS) and network performance.In this thesis, we investigate two types of scenarios: Basic scenario only consists of rechargeable routers; while complex scenario consists of electricity-powered Internet gateways and solar-powered rechargeable routers. The objective is to minimize the deployment and operation costs, under the constraints, including users’ traffic demand, application QoS requirement, nodes’ intrinsic capacity, limited energy supplies and so on. The detailed objectives and constraints are different in the two scenario types. We formulate this placement problem as a mathematical optimization problem, in which we apply an energy flow model for rechargeable routers to characterize the charging and discharging process and propose a new performance metric called network failure rate, instead of the traditional metric of coverage ratio. We also propose cell association algorithms applicable for the two scenarios for connections between mesh clients and mesh routers(and mesh gateways). By incorporating cell association algorithms into classical algorithms searching for optimal solutions, we design feasible and efficient node placement algorithms to obtain near-optimal placement schemes. Finally, a series of simulation experiments are conducted to verify the performance of our algorithms.Specifically,our focus varies according to different types of scenarios. In the basic scenario, we consider to minimize the number of rechargeable routers under constraints including energy, capacity, traffic and so on. Furthermore, we introduce proportional fairness concept which is widely used for resource allocation, and propose two cell association algorithms from different perspectives of efficiency and fairness. While in the complex scenario, we jointly consider the installation capital expenses(CAPEX) and the operational cost(OPEX) with the objective of reducing total costs without compromising user Quality of Service(QoS) and network performance. Two cell association algorithms are proposed to construct links from routers to gateways as well as connections between users and deployed routers.