چكيده به لاتين
Mobile devices’ technology is evolving rapidly. However, their capabilities are not growing at the pace of arrival of new applications, due to their small sizes. As a result, the idea of computation offloading introduced. Based on this idea, users’ applications are sent to a group of powerful servers and resource-rich servers of computation and storage, called Central Clouds. Central Clouds position far from users. So offloading to them will be time consuming and damaging to many applications. Thus, the idea of Edge computing has emerged, which provides computation and storage at the edge of the mobile access networks and has the capability to run applications in their time constraints.
Offloading is a process of hybrid allocation of computational and communicational resources. It means that being provided by one of the resources will not be sufficient to users’ requirements. Moreover, real environments are multi-user, where the users compete to achieve the resources in order that they can conclude their task in their time constraints and also, by use of minimum energy. Another influential factor on computation offlosding decision, is the cost of communicational and computational resources. The decision on computation offloading must be made after considering the amount of resources allocated to the users in accordance with the chosen price. Consequently, considering the importance of hybrid allocation of computational and communicational resources, users’ demands and resource pricing in accordance with it, the necessity of proposing a hybrid and multi user scheme will be declared.
In order to overcome the mentioned challenges, a scheme for computation offloading proposed which uses game theory and optimization, in order to decide on the optimal amount of both resources and optimal unit price of them. This scheme modeled as a Stackelberg game and after proving the uniqueness of the Nash equilibrium point, this point will be computed. Simulation results show that the proposed scheme outperforms the Fixed-Price approach by reducing the amount of unaccepted users due to their disability to pay high prices of communicational and computational resources by 22% and 50% resoectively. It also allocates 40% more communicational resources to users and increases the edge’s utility function by 12% in comparision to the Fixed-price approach.
