Dynamic Task Offloading for Cloud-Assisted Vehicular Edge Computing Networks: A Non-Cooperative Game Theoretic...

Area of Science:

• Vehicular networking
• Edge computing
• Game theory applications

Background:

• Vehicular edge computing (VEC) enhances vehicular networks (VNs) via task offloading.
• Resource-constrained vehicles offload tasks to roadside units (RSUs).
• High mobility and network overload pose challenges for real-time task processing in VEC, degrading performance.

Purpose of the Study:

• To propose an efficient dynamic task offloading approach for VEC to address real-time processing challenges.
• To enhance vehicular performance by minimizing response time and task failure rates.
• To enable vehicles to independently decide optimal offloading strategies to MEC or cloud servers.

Main Methods:

• A non-cooperative game (NGTO) framework is proposed for dynamic task offloading.
• Vehicles employ a best response offloading strategy to achieve a stable equilibrium.
• Task-offloading probabilities are dynamically adjusted to maximize individual vehicle utility.

Main Results:

• The NGTO approach significantly reduces response time and task-failure rates.
• Compared to Local RSU Computing (LRC), reductions were 47.6% and 54.6%, respectively.
• Compared to random and collaborative offloading, reductions were substantial (up to 39.7%).

Conclusions:

• The proposed NGTO scheme effectively addresses VEC challenges related to mobility and overload.
• The strategy ensures performance guarantees by optimizing task offloading decisions.
• This approach offers a stable and efficient solution for improving vehicular network performance.