A New Computing Offload Algorithm for Maximizing User Revenue Under Cloud Edge Architecture

In order to solve the problems of traditional Mobile Cloud Computing mode, such as MCC cloud servers often being far away from users, high network transmission latency, and unable to meet the low latency requirements of application services, the author proposes a new computing offloading algorithm research method that maximizes user benefits under cloud edge architecture. Firstly, the optimization goals of existing work mostly focus on minimizing latency and energy consumption, without paying attention to the benefits of users in the computation offloading process. User benefits are of great significance in guiding users to perform reasonable calculation and uninstallation of their devices: If a user takes a certain calculation action in the current system environment and obtains significant benefits, it indicates that the calculation action is compatible with the current system environment, this will encourage user devices to take the same action choices in similar system environments in the future. Therefore, the author proposes the concept of Total User Revenue (TUR) and introduces TUR into the definition of objective optimization problems. Secondly, the author proposes a new computing offloading algorithm (BDTUR) based on DeepQ Network (DQN) for the computing offloading problem in a multi user, single cell, one cloud, one side vertical collaborative network architecture. This algorithm can make the most suitable offloading decision for the current system environment, and can self-learning based on feedback from the environment, continuously improving the accuracy of the decision. Through simulation experiments, it can be seen that the intelligent unloading algorithm proposed by the author has higher overall user benefits compared to the comparative algorithm.