An Energy-Efficient Data Offloading Strategy for 5G-Enabled Vehicular Edge Computing Networks Using Double Deep Q-Network

In the era of fifth-generation (5G)-enabled vehicular edge computing (VEC), efficient data offloading strategies are essential. The complexities inherent in this environment, such as vehicle speed, direction, position coordinates, signal strength, data to be offloaded, server load, and energy consumption, magnify the need for innovative solutions. Given the intricate nature of this environment, there’s a demand for a novel and robust algorithm to effectively address its complexities. Traditional models and algorithms, while functional, exhibit inefficiencies and high computational demands. This paper presents an innovative data offloading approach, capitalizing on the power of deep reinforcement learning through the implementation of a cutting-edge application known as double deep networks (DDQN). The proposed algorithm merges the capabilities of deep neural networks and techniques from deep learning, facilitating flexible choices that adjust to evolving circumstances. This algorithm incorporates an optimization process within a multi-user context and a sophisticated trade-off analysis, effectively balancing between the different objectives to optimize overall system performance. Our DDQN model, trained on a synthetic dataset, outshines extant methods in key performance indicators. Specifically, we achieved an 80% improvement in energy efficiency, a 72.6% reduction in communication overhead, and a 9.8% improvement in delay. This strategy offers a potent solution for optimizing 5G-enabled VEC, paving the way for enhanced real-time vehicular network services and addressing the challenges of data offloading within this framework.