A GA-Based Strategy for Deploying Cable Connected Roadside Units in VANETs

Although wireless roadside unit (w-RSU) is more flexible than cable connected roadside unit (c-RSU), the excellent performance of c-RSU, such as high communication capability and large communication range, ensures its role in VANETs. Deploying c-RSUs at large area can bring high cable cost and different RSU positions may contribute to distinct cable cost even the number of RSU is fixed, which is often overlooked in current research. Hence, in this paper, we incorporate cable cost into the deployment scheme and treat selecting c-RSU positions as an inseparable procedure with that of selecting road segments to lay cable between c-RSUs. A two layer embedded genetic algorithm (TLEGA) is proposed to make tradeoff between deployment budget and network performance. TLEGA contains an embedded evolution structure so that the RSU position population and road segment population can co-evolve. Simulation results show TLEGA can find optimal deployment schemes within a limited number of run and the advantages of deploying c-RSUs instead of w-RSUs in high frequency zone of accident are also verified.