Design of Distributed Offloading Algorithm Based on Mobile Edge Computing Network

The paper models the interaction process between users and edge cloud based on the Stackelberg game. Firstly, the users' cost is modeled as a minimization problem with the amount of offloaded tasks of each user as a constraint and the revenue of edge cloud is modeled as a maximization problem with the computation capacity of the edge cloud server as a constraint. Then, for the problem of minimizing users' cost, the paper derives the optimal offloading scheme under uniform pricing which follows the threshold strategy and gives the users' bandwidth allocation strategy. Finally, for the problem of maximizing the revenue of edge cloud, the paper substitutes the optimal offloading plan, and finds the optimal price. The paper compares the proposed distributed offloading algorithm with optimized bandwidth allocation with the fully local computing algorithm and the bandwidth sharing algorithm. The average user latency has been reduced by more than 30% and 20%, and the revenue of edge cloud has increased by more than 1.5 times.