Multi-criteria Optimization Approach for the Deployment Planning Problem of Multi-hop Wireless Networks

摘要

The few studies carried out so far on planning Wireless Mesh Networks (WMNs) tend all to be mono- objective optimization models. We propose a different approach to address the planning of WMN problem that reflects as much as possible the real-life problem. Basically, an optimal (or rather a good and realistic) planning solution has to be simultaneously cheap (minimizing the deployment cost) and efficient (maximizing the throughput). To achieve this we devise a novel generic multi-objective optimization model where the two objectives of network cost deployment and network throughput are simultaneously optimized under obvious network constraints. We then derive two instance models differing mainly in how the second objective is defined: maximizing the culmination of the flows over the entire network or minimizing the aggregation of network interferences. Obviously, a new network metric is proposed to handle this task. A comparative experimental study with different key-parameter settings on the two instance models is conducted to help network planner decide which planning optimization model to choose given their specifie requirements and/or scenarios.