N-Deployment: a service-oriented and intelligent infrastructure deployment strategy for vehicular networks in complex urban environments

Abstract

Emerging smart-city applications demand vehicular communication infrastructures that are cost-efficient, scalable, and adaptable to diverse service needs within complex urban traffic environments. This work presents N-Deployment, a flexible strategy for deploying fixed vehicular network infrastructure (RSUs) that meets heterogeneous service requirements. The proposed approach identifies minimal deployment configurations that maximize coverage efficiency and support multiple service types over a unified urban network. N-Deployment integrates insights from prior studies, including infrastructure minimization, geometric optimization, budget constraints, service continuity, unsupervised learning techniques such as clustering, and AI-driven mobility pattern analysis. We formulate a multiobjective optimization model to ensure that a specified fraction of the vehicle fleet connects to the infrastructure within a predefined time window, traversing at least N covered cells along their routes. Additionally, the model enables heterogeneous deployments, delivering differentiated services across distinct urban regions according to local demands. GRASP and NSGA-II metaheuristics are employed to address large-scale scenarios. Experimental results using real mobility traces from Cologne, Germany, show that N-Deployment outperforms traditional strategies by achieving higher vehicle coverage with fewer RSUs while supporting multiple service levels. To our knowledge, this is the first unified strategy to simultaneously address service heterogeneity, infrastructure cost, and real-time connectivity in vehicular network deployments.