Advanced Deep Learning-Based Computational Offloading for Multilevel Vehicular Edge-Cloud Computing Networks

Abstract

The promise of low latency connectivity and efficient bandwidth utilization has driven the recent shift from vehicular cloud computing (VCC) towards vehicular edge computing (VEC). This paper presents an advanced deep learning-based computational offloading algorithm for multilevel vehicular edge-cloud computing networks. To conserve energy and guarantee the efficient utilization of shared resources among multiple vehicles, an integration model of computational offloading, and resource allocation is formulated as a binary optimization problem to minimize the total cost of the entire system in terms of time and energy. However, this problem is considered NP-hard and it is computationally prohibitive to solve this type of problem, particularly for large-scale vehicles, due to the curse-of-dimensionality problem. Therefore, an equivalent reinforcement learning form is generated and we propose a distributed deep learning algorithm to find the near-optimal computational offloading decisions in which a set of deep neural networks are used in parallel. Finally, simulation results show that the proposed algorithm can exhibit fast convergence and significantly reduce the overall consumption of an entire system compared to the benchmark solutions.INDEX TERMS Computation offloading, vehicular edge-cloud computing, autonomous vehicles, 5G, resource allocation, deep reinforcement learning.

A survey on computation resource allocation in IoT enabled vehicular edge computing

“…Power/energy [28][29][30] QoS [33][34][35] Delay [25,26,28,[35][36][37][38][39] QoE and user satisfaction [34,37] Network/system utility [26,33,37,39,40] Reliability [20] However, at the same time, they are required to perform many intensive computations as part of the network [27]. Hence, authors in [28][29][30][31] try to optimize energy and power consumption in their resource allocation schemes.…”

“…Power/energy [28][29][30] QoS [33][34][35] Delay [25,26,28,[35][36][37][38][39] QoE and user satisfaction [34,37] Network/system utility [26,33,37,39,40] Reliability [20] However, at the same time, they are required to perform many intensive computations as part of the network [27]. Hence, authors in [28][29][30][31] try to optimize energy and power consumption in their resource allocation schemes. Optimizing power includes various factors like transmission power, uploading time, offloading ratio, local CPU frequency, etc and all these factors need to be considered while formulating the energy/power optimization problem.…”

“…Their simulation results showed that their joint computation offloading and URLLC resource allocation for collaborative MEC-assisted cellular-V2X networks could improve trade-off performance compared to centralized MEC-assisted V2X. In [28], the authors formulated a problem based on binary optimization problem to minimize the entire cost of the entire system in terms of energy and time and guarantee efficient resource utilization among multiple vehicles which are part of the network [30].…”

Computation offloading strategy based on deep reinforcement learning for connected and autonomous vehicle in vehicular edge computing

“…Although the experimental results show that the performance of the proposed algorithm is better than other benchmark algorithms, the offloading of dependent tasks is not considered. Khayyat et al [18] proposed a distributed deep learning algorithm to optimize the delay and energy consumption. The simulation results show that the algorithm has faster convergence speed.…”

“…1 The optimization objective mainly focused on either execution delay [13,14] or energy consumption [15,16], instead of the joint optimization of offloading failure rate and energy consumption [17,18]. 2 Most work only considered computation offloading strategy in the static environment [19][20][21], instead of dynamic offloading strategy with the change of vehicle positions in different time slots.…”