Game-Theoretic-GAI Approach for Computation Offloading and Resource Management for Mobile Edge Collaborative Vehicular Networks

To counter challenges posed by emerging service needs, limited resources, and the imperative of real-time flexibility, the present research synthesizes game theory and Generative Artificial Intelligence (GAI). The designed architecture supports discreet and scalable resource allocation by combining a Stackelberg game framework with GAI-driven simulation and decision-support mechanisms. The algorithmic methodology of two stages allocates resources among vehicles and Mobile Edge Computing (MEC) servers in an efficient manner, optimizing the offloading ratio. The architecture dynamically adapts by changing network environments using GAI’s ability to simulate complex vehicular scenarios and weight trade-offs concerning latency, energy consumption, and computation efficiency. The resource allocation process is supported by an intelligent offloading decision-support module powered by GAI, complementing parameter optimization using Lagrangian optimization and Karush-Kuhn-Tucker (KKT) conditions. The performance of the designed framework is validated using extensive simulations. The results show the reductions in computational overhead, energy consumption, and latency compared to conventional methods, especially in cases of dynamic job intensity and communication scenarios. The Generative AI can improve the system scalability by allowing task offloading in resource-scarce cases. The result demonstrates the synergistic effectiveness of GAI and game-theoretical approaches in resolving real-time resource allocation challenges in vehicle-to-everything networks.