A spectral clustering-based deployment strategy for roadside units in vehicular edge computing environments

Abstract

Roadside units deployment strategy in vehicular edge computing (VEC) environments plays a crucial role in balancing deployment expenses with user experience. This paper tackles this challenge by formulating the communication and computation cost of offloading vehicle tasks to RSUs and, subsequently, optimizing the RSU deployment problem to minimize the average task completion time. We propose the spectral clustering-based deployment strategy (SCBDS), which determines the optimal number and precise locations of RSUs by identifying clusters of vehicular traffic demands. Performance evaluation through simulations in real-world traffic scenarios demonstrates that SCBDS outperforms existing algorithms in terms of task completion time, showcasing its effectiveness in enhancing VEC efficiency and user experience.