Binary task offloading strategy for cloud robots using improved game theory in cloud-edge collaboration

In this study, we address the challenge of efficiently processing cloud robot tasks within cloud-edge collaborative systems by designing a collaborative offloading system. Our focus lies in jointly optimizing system completion time and energy consumption through well-calibrated trade-offs between the two objectives. To this end, we developed a task-offloading model for binary (indivisible) tasks, fostering cooperation between multiple edge and central cloud servers. We employed game theory to strategize offloading of computing tasks to local devices, edge clouds, or central clouds in multi-robot scenarios. An improved game-theoretic algorithm with polynomial time complexity was proposed, transforming the task-offloading challenge into a game that dynamically updates the offloading strategies and identifies the Nash equilibrium of the system to determine the optimal offloader strategy. Simulation experiments validate the effectiveness of our strategy, demonstrating its ability to reduce energy consumption, shorten the completion time of computational tasks, fully utilize system resources, and significantly improve the quality of cloud-edge collaboration services.